Advice Notes on the Non- Destructive Testing of Road Structures

Transcription

1 Advice Notes on the Non- Destructive Testing of Road Structures AM-STR AM Asset Management & Maintenance Standards

2 TRANSPORT INFRASTRUCTURE IRELAND (TII) PUBLICATIONS About TII Transport Infrastructure Ireland (TII) is responsible for managing and improving the country s national road and light rail networks. About TII Publications TII maintains an online suite of technical publications, which is managed through the TII Publications website. The contents of TII Publications is clearly split into Standards and Technical documentation. All documentation for implementation on TII schemes is collectively referred to as TII Publications (Standards), and all other documentation within the system is collectively referred to as TII Publications (Technical). This system replaces the NRA Design Manual for Roads and Bridges (NRA DMRB) and the NRA Manual of Contract Documents for Road Works (NRA MCDRW). Document Attributes Each document within TII Publications has a range of attributes associated with it, which allows for efficient access and retrieval of the document from the website. These attributes are also contained on the inside cover of each current document, for reference. For migration of documents from the NRA and RPA to the new system, each current document was assigned with new outer front and rear covers. Apart from the covers, and inside cover pages, the documents contain the same information as previously within the NRA or RPA systems, including historical references such as those contained within NRA DMRB and NRA MCDRW. Document Attributes TII Publication Title TII Publication Number Advice Notes on the Non-Destructive Testing of Road Structures AM-STR Activity Asset Management & Standards Document Set Maintenance (AM) Stream Structures (STR) Publication Date Document Number Historical Reference NRA BA 86 NRA DMRB and MCDRW References For all documents that existed within the NRA DMRB or the NRA MCDRW prior to the launch of TII Publications, the NRA document reference used previously is listed above under historical reference. The TII Publication Number also shown above now supersedes this historical reference. All historical references within this document are deemed to be replaced by the TII Publication Number. For the equivalent TII Publication Number for all other historical references contained within this document, please refer to the TII Publications website.

3 Volume 3 Section 1 Part 4 NRA BA 86/14 Advice Notes on the Non- Destructive Testing of Road Structures St. Martin s House, Waterloo Road, Dublin 4. Tel: Fax info@nra.ie Web :

4 Summary: These advice notes publicise the outcome of research in NDT carried out by the Highways Agency in the UK. They do not endeavour to provide full details of techniques in common use, but the full range of available techniques is summarised. Published by National Roads Authority, Dublin 2014

5 DESIGN MANUAL FOR ROADS AND BRIDGES VOLUME 3 ROAD STRUCTURES - INSPECTION AND MAINTENANCE SECTION 1 INSPECTION PART 4 BA 86/14 ADVICE NOTES ON THE NON-DESTRUCTIVE TESTING OF ROAD STRUCTURES Contents First Tier General Advice Note 1 General Guidance Second Tier Areas of Application Advice Note 2.1 Advice Note 2.2 Advice Note 2.3 Advice Note 2.4 Assessing the Conditions in Grouted Ducts in Post-Tensioned Concrete Surveying the Structure of Masonry Arch Bridges Testing and Monitoring the Condition of Concrete Structures Testing and Monitoring the Condition of Metal Structures Third Tier NDT Techniques Advice Note 3.1 Advice Note 3.2 Advice Note 3.3 Advice Note 3.4 Advice Note 3.5 Advice Note 3.6 Impact-Echo (I-E) Sonic Transmission and Tomography for Masonry Bridges Ultrasonic Transmission and Tomography for Post-tensioned Concrete Bridges Electrical Conductivity Ground Penetrating Radar (GPR) Acoustic Emission (AE) Enquiries i

6 ADVICE NOTE 1 GENERAL GUIDANCE Contents Chapter 1. The Role of Non-Destructive Testing (NDT) 2. Scope 3. Format of Advice Notes 4. How to Use the Advice Notes 5. Commissioning and Specification of NDT AN.1-1

7 1. THE ROLE OF NON-DESTRUCTIVE TESTING (NDT) 1.1 Typical uses of NDT for Road Structures are as follows: Gaining increased assurance as to the quality of new construction such as weld testing of steelwork. Gaining increased assurance as to the integrity of earlier construction where similar types of construction have been shown to be defective, such as detecting voids in post-tensioned concrete. Providing an indication of the possible internal structure of old bridges for assessment purposes, such as the structural form of masonry arches. Providing supplementary indications of the condition of structures which are indicating signs of distress, such as the degree of contamination and corrosion of reinforced concrete bridges exposed to chloride attack. Recording acoustic emissions from bearings exhibiting cracking or excessive friction or from bedding to bearings undergoing cracking. The expert interpretation of such emissions can be of assistance with structural appraisal. 1.2 Non-destructive testing is generally carried out by operating equipment close to, against or fixed to the surface of the structure, and has a major advantage that it does not damage the structure. This avoids weakening what may be an already damaged structure. 1.3 In certain circumstances non-destructive testing can be quicker than intrusive investigations and this can reduce costs, such as for traffic management, during the tests. 1.4 Speed and ease of use of NDT techniques can be illustrated by the use of Radar which can be scanned quickly across a potentially voided structure. The validity of the survey can then be confirmed by a more detailed examination of suspect areas by using other tests, such as Impact Echo. A further illustration is the use of Acoustic Transmission which can be used to examine long lengths of parapet rails for potential corrosion. This can then be followed up with a more detailed examination of suspect areas, such as with the use of drilling and direct measurement and an ultrasonic technique to determine the remaining uncorroded wall thickness. 1.5 By its nature NDT is an indirect way of measuring physical features, often relying on the time it takes a signal to reflect off a discontinuity within the structure. The velocity of this signal is dependent upon a property of the structural material which is not known with certainty. 1.6 Expertise and experience is, therefore, required to interpret the data collected and judge what they may mean in terms of physical features or material properties. 1.7 NDT techniques thus do not give definitive answers and should be used to build up an overall picture of what is happening to the structure along with other information. 1.8 It is also generally necessary to carry out selected intrusive investigations of the structure in order to calibrate NDT findings. AN.1-2

8 2. SCOPE 2.1 This series of Advice Notes does not endeavour to provide full details of techniques in common use. However, the full range of available techniques is summarized within the Second Tier documents in order to demonstrate the choice open to the Engineer. The principal objective behind these Advice Notes is to publicise the outcome of the latest research, particularly that commissioned by the Highways Agency in the UK. As further research is completed, it is expected that the Advice Notes will be updated and extended. The documents will seek to provide sources of further information where possible. AN.1-3

9 3. FORMAT OF ADVICE NOTES 3.1 This series of Advice Notes comprises three Tiers. The First Tier is a General Document which describes the role of NDT, explains the format of the Advice Notes and how to use them, and gives advice on Commissioning and Specifying NDT for Road Structures. 3.2 The Second Tier comprises sections devoted to different areas of application, each one summarising the full range of tests available to the engineer. This provides the essential background against which particular NDT techniques may be evaluated. The Second Tier documents drafted to date cover the following areas of application: 2.1 Assessing the conditions in grouted ducts in post-tensioned concrete; 2.2 Surveying the structure of masonry arch bridges; 2.3 Testing and monitoring of the condition of concrete structures; 2.4 Testing and monitoring of the condition of metallic structures. 3.3 Areas that may be considered in the future include: measuring and monitoring strain movements in structures; bridge scour; measuring and monitoring parameters influencing the safe load capacity of structures; construction compliance testing; testing for maintenance of ancillary items. This covers parapets, bearings and their bedding, proprietary joints, surfacing and lighting columns. 3.4 The Third Tier consists of documents covering the different NDT techniques, describing the use of the equipment and the capabilities and limitations of the techniques and their relevance for different applications. The different NDT techniques addressed to date are: 3.1 Impact-echo; 3.2 Sonic transmission and tomography for masonry bridges; 3.3 Ultrasonic transmission and tomography for post-tensioned concrete bridges; 3.4 Electrical conductivity; 3.5 Ground penetrating radar (GPR); 3.6 Acoustic emission. 3.5 A family tree showing the format of the Advice Notes follows this page. AN.1-4

10 Figure 3.1 Format of Advice Notes AN.1-5

11 4. HOW TO USE ADVICE NOTES 4.1 The First Tier document, in describing the role of NDT indicates the type of problems which can be addressed by NDT. 4.2 If it appears that NDT can assist the user with a type of problem, then reference can be made to the Second Tier section relevant to the application in question. This will show the types of information that NDT can provide and whether it is likely to be useful. If NDT can be used to provide relevant information, then reference can be made to the tables of various NDT techniques that can be used. The tables show which techniques are in common use, which are at development stage, which are more economical to use, and which give more detailed information. 4.3 Having identified NDT techniques which the user considers might be appropriate for the problems with his or her application, the user is then referred to the relevant Third Tier documents to learn in detail about the techniques, how the equipment is used and the benefits and the limitations. A final choice can then be made on the technique or techniques to be used, on the extent of the survey and the detail required. 4.4 The First Tier document can then be used to provide guidance on Commissioning and Specification, augmented by additional guidance relevant to the particular techniques chosen, obtained from the Third Tier documents. It is preferable to proceed cautiously and not commission too wide a survey until it has been established that the chosen techniques are proving useful. If they are not, it maybe that the circumstances are not suitable for that technique. That does not mean that the technique would not have been useful in more appropriate circumstances. AN.1-6

12 5. COMMISSIONING AND SPECIFICATION OF NDT Introduction 5.1 The following are general requirements for the commissioning and specification of NDT. The specific requirements for individual test techniques are given in the Third Tier documents. Testing Organisation 5.2 Testing organisations should have an established reputation for carrying out non-destructive testing of the bridge construction type involved. They should have an acceptable Quality Assurance system. They should also have an acceptable Health and Safety record, in particular with regard to working at height, working under The Chemical Act (2008) and the Chemical (Amendment) Act 2010,, working adjacent to live traffic, working with Mobile Elevated Working Platforms, working with electrical equipment and, where applicable, working over water and/or under railway safety requirements. 5.3 Testing organisations should be able to provide a track record of experience in using the intended equipment with an operator experienced in interpreting the results in relation to the structure of, and faults in, the bridge construction type involved. Information Supplied to Tenderers 5.4 The tender documents submitted to testing organisations should require the following information to be provided. (Note the additional requirements contained in the Third Tier documents). location of structure; as-built drawings providing general arrangement, structure and materials of the bridge; elements of the bridge to be tested; whether the structure of the bridge is to be determined or confirmed; whether faults in the bridge are to be detected and/or known faults are to be investigated; results of previous testing such as drilling-in so that the testing organisation can use these for calibration purposes; information on any known faults; preliminary hazard information such as adjacent roads, railway lines, water courses or overhead electrical equipment. (The testing organisation should be asked to confirm all hazards for themselves and carry out risk assessments.) 5.5 The deliverables from the testing organisations should be stated. The report requirements as given below including data given in the following formats: raw and filtered data (or traces); contoured plots where applicable; diagrammatic type interpretation; interpretation of the data. AN.1-7

13 Consistency of Bids 5.6 A pricing document should be drawn up giving: mobilisation costs; cost of test and interpretation at each test location. (Different costs for different degrees of access difficulty); cost of report. Information Required of Tenderers 5.7 The Tenderers should be asked to state the following: (Note the additional requirements contained in the Third Tier documents) their proposed method of access; details of the equipment they will use, including the trade name, model and, where appropriate, frequency of the instrumentation, and similar details of any interpretative software; the method of marking the test locations or movement of the equipment across the structural faces and how this will be indicated on the raw data outputs and on the interpretative drawings; the names, qualifications and experience of the operatives and supervisory staff they will use; Risk Assessments; whether there is a possibility of their equipment interfering with railway apparatus. full details and costs of any calibration drilling or sampling considered necessary for each individual test for accurate interpretation of the NDT. Evaluation of Tenders 5.8 Tenders should be evaluated on a combination of quality and cost, using a dual envelope system whereby the quality bids are first opened and only bids of acceptable quality have their cost envelopes opened. quality marks, at least 60%, should be awarded for: - testing organisation experience of NDT on the construction type involved; - operator experience on the construction type involved with the technique(s) to be used; - methodology adopted; - Health and Safety record; costs marks, should be awarded for: - overall cost; - sensitivity study of cost in event of additional tests being required; - adequate allocation to interpretation and reporting. AN.1-8

14 Note to Commissioning Engineer 5.9 The work on site should be supervised and the operator interviewed so that he or she can demonstrate his or her competence. Also, additional tests may be required in order to determine the extent of structural features or faults. It is important that the work be staged so that later tests are targeted to determine more detailed information about features identified in earlier tests The testing organisation should be required to demonstrate it has carried out Health and Safety Risk Assessments, and taken appropriate action, before starting work, and for any work instructed during the contract which extends areas of investigation. How to Use the Results 5.11 The limitations of the technique must be taken into account in assessing the results Consideration should be given to the use of methods such as drilling into the structure (taking care not to weaken the structural members) to confirm the presence or absence of features or faults at selected critical locations, as NDT techniques do not always provide unambiguous definitive results. In the case of Post-Tensioned Ducts care should be taken not to damage the tendons. NDT Reports 5.13 The Specification should require the following to be included: (Note the additional requirements contained in the Third Tier documents) i) the name and precise location of the bridge using Ordnance Survey co-ordinates; ii) iii) iv) the date, time and weather, including temperature when the survey was undertaken; the dimensions of the structural elements; the location and size of the test areas including orientation relative to North; v) the surveying method chosen for the accurate positioning of the equipment on the structure, and how this is indicated, both on the raw data outputs and on the interpretative drawings; vi) vii) plots of the measuring grids; the trade name, model and frequency of standard instrumentation used, specification of nonstandard instrumentation and similar details of any interpretative software; viii) the type of survey carried out, a clear description of the methodologies of collection of the data, the test locations used and a demonstration as to whether these were adequate to locate structural features, faults or voids at all locations required; ix) the personnel involved in the survey, including their qualifications and experience; x) raw and, if applicable, filtered data to provide a clear view of the quality of the data print-out in a format such that it can be interpreted by a second opinion; xi) a clear indication of the accuracy and sensitivity of the survey based upon the choice of the hardware provided in a format so that the specifying engineer can compare the claimed accuracy with the plots and further interpretation carried out at a later date; AN.1-9

15 xii) a plot of results in tabular and graphical format and an interpretation of these; xiii) drawings showing the positions of structural elements and of clear voiding or faults, together with size or volume if possible, of possible voiding, and of other defects, together with their location and extent; xiv) an indication of the accuracy of the information provided in (xiii) above, and the likelihood of there being unidentified defects. A discussion of the number of readings and a demonstration of whether these were adequate to locate structural features, faults or voids at all locations required; xv) the results of complementary NDT or in-situ tests, such as confirmation of the presence or absence of voids by physical drilling and endoscopic inspection; xvi) photographs of structure including close-ups of areas investigated; xvii) a reference to this Advice Note The following notes apply to the report It is essential that the testing organisation provides not only an interpretation of the cross sections in relation to the internal structure of the bridge but also the actual raw data so that the specifying engineer has a clear view of the quality of the data print-out. The results from a survey must be presented in a format which can be clearly understood and also clearly and unambiguously related to the internal structure of the bridge. Thus, the test locations or traverse lines across the structure should be clearly marked and, in addition, markers should be shown on the records or traces indicating lateral locations where apparatus is moved across the section of the structure under investigation. A diagrammatic interpretation of the results should be provided. Certification Procedure 5.16 The Contractor should be required to provide a certificate signed by a Technical Director of the company confirming that quality assurance procedures have been followed, that the brief has been followed as stated in the report, that the experience and qualifications of the personnel carrying out and interpreting the survey are as stated in the report, and that all reasonable professional skill and care has been used in carrying out the survey, interpreting the results, identifying the structural form and defects, and considering whether further defects remain unidentified. Records of Use of the Techniques 5.17 Until further notice, there will be a need to record the use of these methods on National Roads Authority structures with comments and conclusions by the commissioning body. A form is provided as Table 5.1. AN.1-10

16 Table 5.1 Records of Use of the Techniques Technique/ Equipment Structural Element Dimensions of Element Form of Element Materials of Element Structural Dimension Void Size Cracking Detected Comments AN.1-11

17 ADVICE NOTE 2.1 ASSESSING THE CONDITIONS IN GROUTED DUCTS IN POST- TENSIONED CONCRETE Contents Chapter 1. Background to this area of NDT Application 2. Formulating a Test Programme 3. Potential Testing Techniques 4. Selection of the Most Appropriate Techniques 5. References AN.2.1-1

18 1. BACKGROUND TO THIS AREA OF NDT APPLICATION 1.1 Grouted post-tensioned concrete is a form of construction introduced to Ireland in the mid-twentieth century. Durability has been found inadequate in a number of such structures, leading to a need for managers to undertake special inspections to ensure safety. This document provides further information on test techniques to be considered within such inspections. The document does not cover testing and monitoring that may be used during subsequent management of such structures, testing of new grouting or remedial grouting, or testing of a more general nature appropriate for the reinforced concrete that surrounds grouted ducts. 1.2 Grouted post-tensioned concrete is constructed in stages. Tendons of steel wire or strand are first contained within preformed ducts within the cured concrete. These tendons are then stressed by a jack reacting against the concrete, before they are anchored by wedges within end blocks. When the stressing jacks are released, pre-compression of the concrete thus remains. Cementitious grout is then pressure injected around the tendons. If this grout is not of good quality and does not fully encapsulate the tendons, there is therefore a risk of corrosion to these vital elements, and a reduction in the ultimate capacity provided by the bonding. 1.3 There are now a number of grouted post-tensioned concrete structures on the Motorway and National Road network. At the time this form of construction was first introduced, it was considered economic for spans as small as 15m. When casting yards for pre-tensioned concrete were built, precast pretensioned beams became more economic for those spans up to about 28m that could be transported to site. In more recent years, applications for post-tensioned concrete became largely confined to larger spans, including the use of segmental construction. 1.4 The concern over the durability of tendons in post-tensioned concrete is now worldwide. Surveys on a sample of road bridges undertaken in the 1970s in the UK demonstrated that voids within grout were not uncommon, although risks of collapse were considered at that time to be small. In 1985 Ynys-y-Gwas Bridge in Wales, which was of segmental construction, did collapse without warning, and serious defects were discovered in a number of Motorway and National Road structures throughout the UK. In 1992 a major structure in Belgium collapsed causing loss of life. 1.5 As a consequence of durability concerns, the Highways Agency in the UK instigated a major programme of special inspections of post-tensioned structures. The results of these inspections, to date, have confirmed the early impressions that voids are very much more common than resulting corrosion. 1.6 Non-destructive testing, therefore, has a major role to play in evaluating the condition of the posttensioned bridge stock and its consequences. Bridges of this form of construction are likely to require ongoing evaluation, and their conditions must be well understood if bridge managers are to minimise the cost of needless repair and replacement whilst maintaining public safety. 1.7 Some bridges have been replaced due to concern over the durability of tendons, but most bridges inspected to date are not considered to be at risk. Between these two extremes are a minority of bridges where the investigation has not been conclusive or early replacement or strengthening is not practicable. These structures are at too great a risk to leave in service without further measures, but do not justify closure if safety can be otherwise assured. Monitoring can fill that role. AN.2.1-2

19 2. FORMULATING A TEST PROGRAMME Maintenance Activities that May Require Support by Testing 2.1 As described within Chapter 1, Special Inspections of the conditions within grouted post- tensioned structures are required by most bridge owners. In the case of the National Roads Authority, requirements are specified within NRA BA 50, Post- tensioned Concrete Bridges. Planning, Organisation and Methods for carrying out Special Inspections. The National Roads Authority standard NRA BA 50 outlines the test methods available when the programme commenced, and remains valid. The current Advice Notes on NDT supplement this information, particularly in respect of methods that have been the subject of ongoing research. 2.2 Testing of conditions within grouted ducts in post-tensioned concrete is therefore required to support the following maintenance activities: to assess the current capacity of the structure (i.e. have any tendons failed); to assess the likely durability of the structure (i.e. how likely is future tendon failure); to assess the feasibility and planning of particular remedial works (e.g. re-grouting). Information Required from a Test Programme 2.3 It may be seen from the above that the final goals of the maintenance engineer are centred around tendon corrosion, either at the present time or in the future. The presence of voids within the grout is an important indicator of the probability of poor durability, to be assessed alongside other indicators of potential corrosion. 2.4 The most important benefit of grout is the protection of metal tendons against corrosion, resulting, in particular, from the alkaline environment it introduces. This protection is threatened by ingress of chlorides, and/or by reduction in the alkalinity arising from carbonation. Voids in the grout therefore have two important effects on the protection. Firstly, they demonstrate the full or partial absence of the protective grout, and secondly they facilitate transport of water, contaminants, carbon dioxide, and oxygen which support either carbonation or the corrosion process. 2.5 A further benefit of grout is the resulting increase in the ultimate load capacity of the structure, and the potential for tendons to re-anchor in the event of localised tendon failure. However, this is of secondary importance, because re-anchoring only becomes significant if the structure is overloaded or tendons have already corroded. 2.6 The degree to which voids are sealed has an important influence upon the long term durability of tendons. Corrosive conditions are much more likely if voids are not sealed, and chlorides and/or air can gain access. Indeed, a very small void to which chlorides can gain ingress is a high risk, whereas a very large sealed void is low risk providing a small amount of grout has created alkaline conditions. The protection afforded by all grout will break down slowly if carbon dioxide is able to promote carbonation. Corrosive conditions may be established relatively quickly if water contaminated by chlorides can gain access. AN.2.1-3

20 2.7 The information required by the maintenance engineer from a test programme should therefore include the following: the location of the ducts and tendons. (To check construction tolerance, but in particular to avoid damage if intrusive investigation is required); the presence and characteristics of voids within ducts. (If the duct has voids, alkalinity will be reduced, contamination can spread, and bond/ anchorage are reduced); the existing degree of tendon corrosion. (Are the tendons still stressed?); the probability of future corrosion. (This depends upon alkalinity and contamination, together with the probability of these changing due to inadequate encapsulation and leakage) 2.8 The bridge engineer is likely to obtain the evidence he requires by combining a number of different examination and test methods, each one offering particular advantages and disadvantages: external visual examination. (Essential evidence, but not adequate on its own); NDT external examination. (No structural damage, but methods vary in reliability and accuracy); intrusive drilling and examination. (Unambiguous localised knowledge of internal conditions, but requires accurate and careful drilling); pressure testing within intrusive drilling. (Extends knowledge of defects and conditions, but relies upon intersection of a void.) 2.9 The advantages and disadvantages of different test techniques are summarised within Chapter 3. The set of information gathered for a given structure will inevitably be a compromise judgement between a full set of information on the one hand, and on the other hand the costs and risks of damage required to gather that information The maintenance engineer will therefore make his judgement by combining samples of the following data: the characteristics of voids in the grouting. (Location, frequency, continuity, tendon protection, etc.); existing duct conditions. (Carbonation, chloride content, tendon conditions); the degree of tendon encapsulation. (Degree of duct sealing, location of potential contaminants) Monitoring may be considered as testing over an extended term. For certain applications, intermittent measurements at widely spaced intervals are effective, as conclusions may be drawn from overall trends in say strain or half-cell potential. Unfortunately such measurements are not effective for monitoring on-going corrosion of prestressing tendons Failure of prestressing tendons will change the strain of concrete close to the point of failure, and over a wider area if the tendon is unbonded due to very poor grouting. Vibrating wire strain gauges programmed to take readings at short time intervals have detected such failures, but to do so reliably they must be located close to the point of failure. Optical fibre sensors can measure strain over a longer length, but are then less able to identify the precise location of the failure. It has also been found that the strains can also return to their original values as the concrete reacts to the localised event. Strain monitoring can therefore be used to monitor loss of prestress, but this requires sensors at all the points at risk, and frequent measurements. AN.2.1-4

21 2.13 A more efficient form of on-going monitoring of post-tensioned concrete is acoustic emission. This comprises the detection of the small amplitude elastic stress waves caused by tendon fracture. Detection of the high energy released by such fractures is well within the capability of the available equipment, which can detect micro-cracking that is many orders of magnitude smaller. An array of sensors is also able to locate the fracture by analysis of the time differential the signal is received at each sensor. AN.2.1-5

22 3. POTENTIAL TESTING TECHNIQUES Technique Description Comments/Value External Visual Inspection Internal Visual Inspection Void Volume Measurement External examination for cracking and deflection. Inspection within drilled holes with borescope, and perhaps with endoscope for very large voids. Volume measurement from drilled holes. Essential, but unlikely to demonstrate internal conditions. Requires care to avoid damage, and results localised, but provides reliable localised information. Volume indicates void extent, but may not indicate corrosivity of conditions. Pressure Testing Measurement of void continuity and void leakage. Leakage provides an indication of corrosivity within voids. Material Sampling Tendon or Concrete Stress Testing Covermeters Radiography Ground Penetrating Radar (GPR) Impact Echo Ultrasonic Transmission/ Tomography Sampling of duct grout, contained water and cover concrete. Measurement of stress within tendons or surrounding concrete by stress relief methods. Covermeters, both standard and adapted for use down drilled holes. Radiographic images of ducts and reinforcement from isotopes or x-rays. Electromagnetic echo sounding method in which transmitter/receiver is passed over the surface at controlled speed. Radio energy pulses are reflected from material boundaries and features, with different electrical permittivity, to formulate a continuous cross-section. A stress pulse induced by impact is reflected at material boundaries, where different acoustic impedances exist, and detected by single accelerometer or displacement transducer. Ultrasonic pulses are received on opposite, adjacent or the same face by receivers. Pulse velocity depends on material properties, and pulse travels round voids. Tomography can additionally build up a 3-D image of the internal structure. Analysis of samples can indicate presence and/or probability of corrosive environment. Tendon stress measurement requires good access. Specialist interpretation required, but data of direct relevance. May be used to locate reinforcement and shallow ducts, and down drilled holes may indicate duct and reinforcement position. Potential for accurate images of ducts and voids, but potential limitations due to restricted penetration, congested sections and safety precautions. Low hazard method, the principal use for which is the location of ducts and reinforcement in advance of alternative methods. GPR has been shown to distinguish empty plastic ducts from full ducts but the orientation of the bow-tie antenna is critical. Voids within metal ducts cannot be distinguished. Low hazard method requiring access from one side only, which can detect voids within metal and plastic ducts. However, only voids that present a large target size may be detected. Prior information on construction details required from radar survey and drawings. Expert interpretation off site required. A low hazard method, but the accuracy at detecting voids is unproven. Must be used in conjunction with other methods to plan and interpret investigation. Slow and not appropriate as primary investigation method. May be appropriate for detailed investigation of an area of known defects. Tomography can potentially differentiate duct voiding from concrete honeycombing. Table Test Techniques AN.2.1-6

23 Technique Description Comments/Value External Visual Inspection Periodic Strain Measurement Automated Strain Monitoring Periodic external examination for cracking and deflection. Measurement of strain at widely spaced intervals. Measurement of strain and temperature at short intervals, continuously recorded on a datalogger. Should be undertaken for any structure at risk, but cannot be relied upon to give adequate warning of structural failure in the event of prestress corrosion. Likely to detect only comparatively large prestress losses, and difficult to separate out the effects of structure strains due to temperature. May detect tendon failure providing sufficient strain gauges are placed close to point of failure. Careful post- processing required to highlight significant data. Acoustic Emission Energy released by tendon failure propagates very large (in AE terms) amplitude elastic stress waves which can be detected as small displacements by transducers mounted on the surface. Table Monitoring Techniques Effective means of detecting and locating failure. Results may be transmitted to remote office. Expensive but justified if large structures at risk are thereby kept in service. Summary of Test Techniques External Visual Inspection 3.1 Principles behind technique: Inspection of the structure for signs of cracking staining or deflection that may indicate internal deterioration. 3.2 Equipment Required: Access equipment required for close inspection. Crack gauge to measure crack widths if found. 3.3 Accuracy: N/A. 3.4 Advantages: Low cost, as test may be undertaken as part of standard inspection programme without specialist equipment. 3.5 Disadvantages: Although cracking is likely to occur before collapse, this should not be relied upon to indicate defects in post-tensioned concrete. It is also beneficial for deterioration to be detected at an early stage, so that remedial measures may be properly planned and safety ensured. 3.6 Notes on use of technique: Visual examination is of no additional cost if it forms part of a standard detailed inspection. Close access is required to detect fine cracking. 3.7 Cracking may occur on the surface of the concrete along the line of tendons due to freeze thaw damage of wet ducts. This has identified structures with defective grouting, although the cracks are fine and may only be noticed when enhanced by salt deposits. AN.2.1-7

24 Internal Visual Inspection 3.8 Principles behind technique: viewing internally through holes drilled to intersect the duct. 3.9 Equipment Required: Percussive drill, generally 25mm dia. A borescope is required for close inspection. Endoscope may be used to penetrate very large voids with good access. Camera attachments required for permanent record Accuracy: Reliable and unambiguous information provided local to the drilled hole Advantages: Results do not rely upon uncertain interpretation of data. When combined with pressure testing and sampling techniques, information of direct relevance to corrosive conditions and the extent of voids remote from the drilled hole can also be gathered. Information on grout quality, water analysis and duct contamination available from samples taken within the same access hole Disadvantages: Drilled holes are semi- destructive, and can introduce a more corrosive environment to the tendon unless carefully reinstated. Sample information is restricted to the locality of the drilled hole, which must be accurately drilled to intersect potential voids, generally the top of the duct Notes on use of technique: Internal visual inspection can provide unambiguous information not available by any other means. However, such testing should be undertaken with great care. Ducts and reinforcement should preferably be located before drilling. Among the NDT methods radar is the most promising, although on many structures ducts have been intersected satisfactorily from construction drawings. Templates should be prepared in advance to guide the direction and location of drilling. Electrical devices to halt drilling as soon as contact is made with ducts can minimise damage. Diamond drills should never be used. Water jetting may reduce damage, but water entry to the duct can encourage corrosion. Void Volume Measurement 3.14 Principles behind technique: Volumes of voids within ducts that are connected to drilled holes may be estimated by pressure methods. Equipment may be based upon the Boyles Law principle, or upon the timed leakage of air flow Equipment Required: Specialised equipment is available for attachment to drilled holes based on either of the principles above Accuracy: The accuracy of the Boyles Law apparatus will be influenced by the void volume relative to the apparatus volume and the degree of natural duct leakage. The accuracy of the method based upon timed leakage will be less influenced by natural duct leakage and void volume Advantages: The volume of voids is an important indicator of grouting quality, particularly when information is also recorded on the leakage from the duct and the continuity of voids Disadvantages: Drilled holes intersecting the duct are required. This may not be practicable if the ducts are within deep concrete sections or masked by other ducts or congested reinforcement Notes on use of technique: The methods are particularly useful when combined with other techniques, and become essential if remedial grouting is to be considered. AN.2.1-8

25 Pressure Testing 3.20 Principles behind technique: Pressure is applied at drilled holes, and resulting leakage flowrates and pressures are observed. Leakage at accessible locations may be observed by use of soap solution Equipment Required: Source of compressed air, pressure gauges, and flowmeter. (Preferably electronic) 3.22 Accuracy: Pressures and flowrate may be accurately recorded Advantages: The information on duct leakage is of direct relevance when assessing environmental influences upon the durability of tendons. Critical leakage pathways that risk ingress of de-icing salts can be detected, including those within congested anchorage zones that are extremely difficult by any other means. Void continuity may be indicated by the reaction of pressure gauges, and thus, indicate the likelihood of contamination spread. These secondary factors have a greater influence on durability than the presence of a void Disadvantages: Drilled holes are required, and must intersect the void location, generally at the top of ducts Notes on use of technique: The major risk and expense is the drilling of holes. If this has been done, information should be maximised by these techniques. Material Sampling 3.26 Principles behind technique: Samples of grout and water are extracted from ducts through drilled holes Equipment Required: Sample containers, and a means of extracting the samples through small holes, such as a flexible tube and vacuum device Accuracy: Accurate analysis of alkalinity and contamination levels is possible Advantages: The carbonation and chloride levels measured are of direct relevance when predicting tendon corrosion. The presence of voids within the grout is of secondary importance Disadvantages: Removal of uncontaminated samples requires care Notes on use of technique: Carbonation tests are required without delay. Tendon Stress Testing 3.32 Principles behind technique: The stress remaining in tendons may be measured by instrumented drilling of small indentation in the tendon Equipment Required: This is a specialist service offered by testing firms who also undertake analysis of results Accuracy: Variable accuracy, depending on nature of the tendon steel Advantages: The method measures a parameter of direct relevance when assessing the current condition of a structure, and also provides information on tendon relaxation since construction. AN.2.1-9

26 3.36 Disadvantages: Fairly large access to tendon required to fix strain rosette, e.g. 100 mm dia core Notes on use of technique: These techniques are unlikely to be a standard test within all inspections, but are of particular value if the existing capacity of a structure is of concern. Covermeters 3.38 Principles behind technique: Measurement of the location of reinforcement and shallow ducts with external covermeters, and detection of reinforcement and ducts from within drilled holes Equipment Required: Standard covermeter for external use or specially adapted device for use within 25mm drilled holes Accuracy: Accurate for reinforcement and ducts close to probe Advantages: An economical means of duct location where low cover and uncongested reinforcement enable its use. Use of covermeter adapted for drilled holes can avoid damage to ducts and reinforcement by frequent use as the hole is advanced Disadvantages: Deep ducts and reinforcement only detected by frequent measurement within drilled holes if damage is to be avoided Notes on use of technique: Measurement of reinforcement location is essential before any structure is drilled. A device adapted for use within 25mm drilled holes can indicate the presence of reinforcement close to any part of the hole, and hence, give advance warning before tendons or reinforcement is damaged. Radiography 3.44 Principles behind technique: Imaging of voids and embedded metal by projection of x-rays or gamma rays onto a sensitive film placed behind the concrete member Equipment Required: Radioactive source, such as iridium, or source of x-rays Accuracy: Clear image can be obtained, but images become more difficult to interpret if the sections are thick, or confused by overlapping ducts or reinforcement Advantages: Extent of voids may be seen, and fractured tendons have been located. May sometimes be used whilst a bridge remains open to traffic, and can produce images quickly from a laboratory on site. Under favourable conditions, interpretation is unambiguous Disadvantages: More powerful sources required to penetrate thicker sections, or to obtain real time images, with a consequent increase in cost and strict health and safety precautions. Radioactive sources emitting gamma rays can penetrate a maximum of 150mm for iridium, and 400mm for cobalt 60, and must be mechanically returned to a sheathed box. X-ray sources may penetrate up to 1500mm, and the ability to turn off electrically is a safety benefit Notes on use of technique: When access and safety conditions are favourable, this is a very useful NDT technique providing visual images that are easily interpreted. The number of suppliers of this service has decreased. AN

27 Ground Penetrating Radar (GPR) 3.50 Principles behind technique: GPR is an electromagnetic echo sounding method where a transducer (transmitter/receiver) is passed over the surface at a controlled speed. Short duration pulses of radio energy are transmitted, and the receiver detects reflections from material boundaries and features, with different electrical permittivity, such as embedded metalwork or voids. The collected data is effectively a continuous cross section. The amplitude, phase and velocity are influenced by the material type, and the continuity of the signal influences the shape of the components. The travel time of radio pulses is influenced by the layer thickness or depth to embedded features Equipment Required: A range of different equipment specifications are available, each providing alternative combinations of cost and accuracy, and it is important that these are selected to be suitable for both the structure and its features under investigation Accuracy: The depth of features may be estimated to an accuracy of + or 10% when the pulse velocity is calibrated by measurement of a section of similar concrete of known thickness. Velocity is influenced by variations in moisture and salinity Empty plastic ducts have been distinguished from fully grouted ducts under laboratory conditions where the bow-tie antennae are aligned perpendicular to the long axis of the duct, but GPR cannot generally be used to determine the depth of a void Advantages: Safe to use, as there is no radioactive or x-ray hazard. Particularly suited to detecting changes on an extensive structure with very restricted access opportunities, and for completely nondestructive location of internal features. Can locate metal ducting, reinforcement, and tendons within fully grouted plastic ducts Disadvantages: Data may only be interpreted by a specialist, and processing off site is recommended with consequent delay before use of information. Results are influenced by structural features, which may complicate interpretation and require knowledge of the structure. Cannot penetrate metal to detect voids. Results are insensitive to the depth of voids. Less sensitive to voids projecting a small target size, (e.g. voids of small vertical height viewed horizontally). Congested ducting and reinforcement is also likely to pose problems Notes on use of Technique: Location of metal ducts by this method is well established, and is likely to be the primary application of GPR for NDT of post- tensioned concrete bridges. Accurate location of ducts and reinforcement is essential knowledge for the selection of equipment required for other NDT methods such as Impact Echo and Ultrasonic Transmission, and the interpretation of data. GPR data should, therefore, be processed off site in advance of these methods. GPR may also be useful in advance of drilling to intersect potential voids in metal ducts, and thus, minimise abortive drilling and reduce the risk of damage to tendons Sources of Further Information: Advice Note 3.5 of this series of Advice Notes. Impact Echo 3.58 Principles Behind Technique: A stress (sound) wave is generated by a short duration impact on the structure. These low frequency stress waves are propagated through the structure and reflected by flaws and external surfaces, provided that they are of different acoustic impedance. Surface displacements caused by the arrival of reflected waves at the impact surface are recorded by a transducer. These are processed to display their amplitude and frequency, and may be saved for postprocessing Equipment Required: Ball bearing impactor. Response transducer. Analyser. Display screen. Power supply. AN

28 3.60 Accuracy: The thickness of concrete slabs may be obtained to an accuracy of 3% The accuracy of void detection may be expressed in terms of the minimum lateral target size of voids that may be detected. This varies according to the impactor diameter, which influences the input frequency a smaller impactor gives a higher input frequency, higher resolution, but shallower depth penetration. This is discussed in more detail elsewhere Advantages: Safe to use, as there is no radioactive or x-ray hazard. Can detect the presence and depth to voids within both metallic and plastic ducts, the depth to reinforcement, and thickness of elements. Requires access from one side of element only Disadvantages: The minimum size of detectable voids is above the diameter of many ducts, and above the size at which voids are of significance for their effect on tendon durability. Most ducts are not accessible from the soffit in critical areas, further reducing the likely target size if viewing is constrained from the side. Water filled voids are not detected. No information is provided on structure beyond the initial void surface, although approximate void depth may be deduced from the duct geometry if measurements are from beneath and a horizontal surface is assumed. The method is not as fast as impulse Radar, and requires prior knowledge of duct geometry to facilitate interpretation. Congested ducting and reinforcement are also likely to pose problems Notes on use of Technique: GPR processing required in advance of impact echo testing in order to direct test locations and enable the estimated frequency response to be modelled. Design details showing the shape and size of ducts and tendons are essential for modelling of the impact echo response. Data analysis off site is required for thorough interpretation Sources of Further Information: Advice Note 3.1 of this series of Advice Notes. Ultrasonic Transmission/Tomography 3.66 Principles behind technique: Ultrasonic pulses are emitted by transducers, and received by transducers which may be placed on opposite or adjacent faces in transmission mode, or the same face in reflective mode. These pulses travel at a velocity that is dependent upon the density and elastic properties of the material, and will travel round voids. Comparison of pulse transmission time periods, therefore, indicate alternative routes for the pulse, and the properties of intervening materials. Tomography assembles data for a 3-dimensional image of the internal construction, from which the structure at a specified section or slice may be drawn Equipment Required: Source of ultrasonic pulses, receivers, and data storage device Accuracy: Ultrasonic transmission is well established for finding voids but its application and the use of tomography for locating voids in grouted ducts is at the experimental stage, and accuracy cannot yet be quoted. The presence of voids has been indicated with a grid spacing similar to the void size Advantages: Safe to use, as there is no radioactive or x-ray hazard. In reflective mode, testing may be undertaken from one face. Some equipment may be used without the special couplant required for impact echo transmitters/receivers Disadvantages: Unproven accuracy for void detection. Requires additional information from other methods, such as radar and impact echo, to plan the investigation and interpret results. Pulse velocity is influenced by moisture content. Interpretation is difficult. Tomography in particular must take account of surrounding structural features. Congested ducting and reinforcement are also likely to pose problems. AN

29 3.71 Notes on use of technique: Very slow to use, and not appropriate for a primary detection of voids throughout the length of a duct. May be appropriate for detailed investigation of areas of voiding identified by alternative methods, providing further work demonstrates accuracy. Although direct transmission cannot distinguish between concrete honeycombing and duct voiding, tomography can potentially do so. This is not yet proven Sources of further information: Advice Note 3.3 of this series of Advice Notes. Monitoring by Acoustic Emission 3.73 Principles behind the technique: Energy released by micro-cracking in a structure propagates as small amplitude elastic stress waves or acoustic emissions which can be detected as small displacements by transducers mounted on the surface Equipment required: An array of acoustic Emission Sensors depending on the extent of the structure to be monitored, preferably manufactured under strict quality controls. Low frequency sensors are usually used for monitoring concrete structures. Cable, power supply, technical equipment, computers Accuracy: Systems with sensitivity to detect and locate signals with 1/10,000th of the energy of a 0.5mm pencil lead break (the HSU-Nielsen field calibration standard per BS EN :2000) exceed that required for locating the early stages of concrete microfracture. The effectiveness of source location by direct linear positioning can be confirmed using lead breaks as an artificial source. Tendon fracture releases energy that is orders of magnitude higher, and is, therefore, easily detected Advantages: Powerful technique, well established in the oil industry. Proprietary system using the technique is already in service, and has demonstrated tendon fractures Disadvantages: Expensive, but can be good value Notes on Use of the Technique: Can be used for short or long term monitoring. AN

30 4. SELECTION OF THE MOST APPROPRIATE TECHNIQUES 4.1 Testing programmes will inevitably include a range of tests, and should be devised with the following factors in mind: A balance must be struck between obtaining sufficient information to make a reasonable judgement on risk, and seeking so much information that the examination itself compromises durability by intrusive sampling. Tests may be effective in combination, e.g. ducts may be located by specialist probes or radar in preparation for intrusive drilling and sampling or impact-echo testing. Tests may be interpreted in combination, e.g. a representative sample of locations revealing a particular characteristic may be examined in greater detail by a variety of more detailed tests. Testing programmes can only be provisional, and may require amendment as a result of continuing testing and interpretation. Staged testing, permitting interpretation of results between each stage, may be appropriate on larger jobs. Intrusive drilling and sampling can provide invaluable information on materials, corrosion details, void volumes, and atmospheric leakage. However, both the drilling and the subsequent sealing of the holes must be carefully specified and supervised to avoid potentially serious damage to the structure. 4.2 The current routine procedure in Ireland and the UK for assessing grout integrity is to drill and inspect visually. As this method is subjective and ad hoc in its accuracy, a more systematic NDT method might be preferable. Radar will not penetrate through metallic ducts which exist in most posttensioned bridges constructed before the mid-1990s. Whilst it cannot, therefore, detect internal voids, its primary application is likely to be confirmation of the location of ducts and reinforcement prior to other forms of testing. However, radar will penetrate through plastic ducts and also through unlined cast-in-situ joints in segmental precast post-tensioned bridges and may be able to detect voids in such cases. Radar cannot detect the depth of voids. For metallic ducts, impact-echo is the most developed NDT technique, and has been promoted heavily in the USA. Voids presenting a relatively large target size may be detected. The newest technique is that of ultrasonic tomography which may be capable of providing a contoured section through the element indicating void locations. The method is slow and requires expert interpretation. At a simpler level, duct location may also be indicated by use of covermeter probes inserted into drilled holes. 4.3 Tendon fracture can be registered and located at the time it occurs by acoustic emission monitoring. This technique can also be used at much higher degrees of sensitivity to detect concrete microfracture as it occurs, for example where there may be problems at post- tensioned anchorage blocks. 4.4 It is advisable to proceed cautiously and not investigate too much of a structure until it has been established that the method is producing useful results. The effectiveness of these NDT methods may be heavily influenced by the geometry and materials of the structure under investigation, such as the depth of ducts and the congestion of reinforcement and other ducts. The accuracy of each method in particular circumstances remains unpredictable in the current stage of development. The effectiveness of each NDT method in the prevailing situation should, therefore, be predicted in advance of specification, and the accuracy of this prediction should be checked before reliance is placed on the results. AN

31 4.5 Non-destructive testing techniques are not definitive, and require calibration. These Advice Notes illustrate the need to frequently evaluate the results from different tests in combination, in order to achieve meaningful interpretation. This may require a return to site after analysis of early test data. 4.6 The final objective behind the testing programme is to quantify, as far as practicable, a prediction of strength and durability. A wide range of test methods is available. The factors influencing test selection are outlined above, and the choice is complex. Nearly all structures differ in their suitability for individual methods, whether they are developing NDT or more conventional methods. In each case, the engineer must consider the prevailing circumstances, and in the light of that knowledge, select the combination of tests that together builds up a picture of the structure and its condition. There will always be a need to balance cost, effectiveness, and the potential consequences that depend upon the diagnosis. AN

32 5. REFERENCES 5.1 The National Roads Authority Publications NRA Eirspan Bridge Management System Manuals NRA BA 86 Advice Notes on the Non-Destructive Testing of Road Structures. Advice Note 3.1 Impact-Echo (I-E). Advice Note 3.3 Ultrasonic Transmission and Tomography for Post-tensioned Concrete Bridges. Advice Note 3.5 Ground Penetrating Radar (GPR). NRA BD 27 Materials for the Repair of Concrete Road Structures NRA BA 35 Inspection and Repair of Concrete Road Structures. NRA BA 50 Post-tensioned Concrete Bridges. Planning, Organisation and Methods for Carrying Out Special Inspections. NRA BD 54 Post-tensioned Concrete Bridges. Prioritisation of Special Inspections. 5.2 BSI Publications BS 1881: Part 201: Guide to the use of non-destructive methods of testing hardened concrete. BS 1881: Part 205: Recommendations for the radiography of concrete. 5.3 Other Publications Guide to testing and monitoring the durability of concrete structures. Concrete Bridge Development Group Technical Guide no 2. March Durable Post-Tensioned Concrete Structures. Concrete Society Technical Report 72 (2010) Concrete Society Non-structural Cracking of Concrete 4 th edition (2010) Woodward, R.J. and Williams, F.W. (1988) Collapse of Ynys-y-Gwas Bridge, West Glamorgan, Proc. Institution of Civil Engineers, Part 1: Vol 84, August, Woodward, R.J. and Williams, D.L.S. (1991) Deformation of segmental post-tensioned concrete bridges as a result of corrosion of the tendons, Proc. Institution of Civil Engineers, Part 1: Vol 90, April, AN

33 ADVICE NOTE 2.2 SURVEYING THE STRUCTURE OF MASONRY ARCH BRIDGES Contents Chapter 1. Background to this area of NDT Application 2. Formulating a Test Programme 3. Potential Testing Techniques 4. Selection of the Most Appropriate Techniques 5. References 6. Appendix A Details Needed for Load Assessment AN.2.2-1

34 1. BACKGROUND TO THIS AREA OF NDT APPLICATION 1.1 Masonry arch bridges form a critical part of the transportation system in Ireland, since they comprise over 80 per cent of the bridge stock in current use. In total, there are over 25,000 masonry arch bridges in Ireland over 1.8m length, with approximately 600 masonry arch bridges on the National Roads Authority s road network. Many road bridges were originally designed for horse drawn traffic and although they are carrying loads greatly in excess of those for which they were designed, they are showing little sign of distress. 1.2 However, within the European Union, a special problem arises with respect to masonry arch bridges due to the increasing axle loads dictated by EU policies. Implementation of these policies in Ireland and the UK means that vehicles with 11.5 tonne axle loadings now run on Irish and UK roads. The increased traffic loading since the original design and construction of many of these bridges is one of the main causes of concern, together with material deterioration and the development of structural faults resulting from inadequate specification of the construction method and the materials used. The assessment and maintenance of the structures is a difficult and expensive task, especially while the bridge is in service. 1.3 Many masonry arch bridges are historic structures and rarely have accurate or perhaps any drawings of their construction. Considerable problems occur when it becomes necessary to assess the structural stability of a critical component of their complex structure. Hence, it is essential that some knowledge of the internal structure of a bridge is obtained before certain forms of remedial work or strengthening can be carried out. In this respect, Non-Destructive Testing (NDT) methods can play an important role both in the inspection process and at a later date when the outcome of the strengthening process has to be checked. 1.4 Because of the difficulty in determining the quality of the structural materials and the condition of the structural elements, and identifying hidden features, idealisation of old arch bridges can never be very accurate without substantive investigation of the existing structure. Non-Destructive Testing can be an effective tool for identifying the properties of these bridges without causing damage to the structure or disruption to road users. AN.2.2-2

35 2. FORMULATING A TEST PROGRAMME Maintenance Activities that may Require Support by Testing 2.1 The reasons for surveying the structure of a masonry arch bridge may be broadly categorised as follows: assessment of the existing load capacity; strengthening, widening or other modification to the structure; investigation to determine the causes of irregularities identified during an inspection, which may affect the strength or durability of the structure. 2.2 NDT has a role to play in each of these maintenance activities, but it must be focused towards meeting the specific requirements of the maintenance engineer. Information Required from a Test Programme 2.3 Most arch bridges do not have reliable records of construction or early repair details. It may, therefore, be difficult to determine the physical dimensions of the main structural elements, or the presence of features such as haunching (backing) at supports, saddles over the barrel, internal spandrel walls and ribs. (See figure 2.1). Many unsuspected examples of these features have only been discovered during demolition or repair. Inadequate knowledge of the layout and dimensions of the structural elements may compromise all analysis, repair and strengthening. A summary is therefore given of the information about masonry structures most likely to be required by the engineer. Figure 2.1 Examples of Sections through bridges which have revealed hidden features at demolition AN.2.2-3

36 2.4 Specific information to be obtained by testing for each maintenance activity is summarised as follows: Assessment of Load Capacity 2.5 The assessment of load capacity will require some knowledge or estimation of the following (see Appendix A): the geometry of the arch survey from beneath (not NDT); the arch barrel construction thickness, integrity and material properties (potential NDT); the backfill fill depth and material properties (potential NDT); hidden features ribs, chambers etc. (potential NDT). 2.6 Assessment by calculation can be imprecise, given the variable properties of the materials, uncertain dimensions, and the simplification required to represent the structure during analysis. Assessment engineers are reassured by the fact that cracking or deformation can be expected before failure, rather than sudden changes such as might be caused by scour. Engineers may be less concerned by the results of such assessment calculations if a structure is carrying full load, shows no signs of distress, and is believed to be in a stable condition. 2.7 Stability of load-critical features thus becomes a key issue. Cracks that are essentially perpendicular to the surface can be monitored by inspection or simple instrumentation. Typical examples of these are the very common cracks in the soffit, parallel to the face ring that indicates either a spandrel wall moving outwards, or differential/concentrated loading on the top of the barrel. A significantly more serious consequence may arise from cracks which are in the same plane as the barrel, such as occur with delamination between successive arch rings. This delamination reduces the effective thickness of the barrel, and can deteriorate particularly in the presence of water, without being evident from visual inspection. The defect is usually first discovered during inspections by drumminess (a hollow sound identified when the surface is routinely tapped with a hammer). More sophisticated techniques may offer more reliable information, and also verify the efficiency of repairs. Strengthening or Widening of a Structure 2.8 Strengthening or widening of a structure may comprise a wide range of interventions, each of which must be tailored to suit the geometry of a particular structure. The most common interventions are as follows: exposing the barrel to add a saddle and/or waterproofing; removal of fill to add a high-level slab, perhaps also tied to spandrel walls and waterproofed; addition of tie bars and patress plates between spandrel and/or wing walls; grouting and closely spaced steel pin stitching to consolidate delaminated areas of arch barrel; strengthening of the arch barrel by steel reinforcement; widening of the structure with additional foundations and spandrel wall; it is particularly important to understand the extent of works before the contract starts on site, because discovery of unexpected features can have serious contractual implications and cause network disruption. The means by which this information is obtained will depend upon the relative cost and reliability of each potential method, when applied to the particular structure under consideration. AN.2.2-4

37 Investigation to Explain the Causes of Irregularities 2.9 Irregularities discovered during inspection may be sub-divided into two main groups: either those that relate to physical movement of an element; or the degradation of construction materials Irregularities that relate to physical movement may include the following: deformation of barrel soffits or walls; settlement of foundations; settlement of surfacing or fill material Effective remedial measures depend upon some understanding of the causes of the problem. NDT may contribute information that increases understanding of the causes, although these will frequently be self- evident Irregularities that may be described as material degradation include eroded mortar, cracked blocks, water erosion, frost or chemical attack. The reason behind the degradation must be understood, although both the cause and the appropriate repair will usually be clear without further testing. If testing is required, material sampling is most likely to provide the information to enable a forecast of future performance to be made. AN.2.2-5

38 3. POTENTIAL TESTING TECHNIQUES Technique Description Comments/Value Visual Inspection and Hammer Testing Coring Trial Pits Sonic Transmission Sonic Tomography Electrical Conductivity Ground Penetrating Radar Infra-red Thermography External examination of surface appearance. Hollowness indicates extent of any de-lamination. Holes drilled into structure to identify materials and condition. Holes excavated into structure to identify materials and condition. Measurement of stress wave transmitted through structure. As above, except section crossed by a net of wave paths. Small currents induced in structure and sensed by a receiver coil. Electromagnetic impulses transmitted into structure and recorded by a receiver. Thermal imaging camera. Currently under research. Essential. Quick; low cost. Identifies surface information. Unlikely to demonstrate internal conditions. Provides reliable localised information. Provides reliable localised information. Can provide comprehensive coverage and good depth penetration. Slow. Does not define depth. Provides increased definition from sonic transmission. Enables an estimation of the internal structure to be made. Provides rapid, comprehensive coverage. Data needs to be used in conjunction with other tests. Will detect the presence of metal. Gives fairly rapid and comprehensive coverage. Cannot penetrate conductors such as metal. Very rapid non-contacting technique; no access required. These can be used to calibrate NDT Techniques. Endoscopes can be used in cored or drilled holes. Table Test Techniques Summary of NDT Techniques Visual Inspection and Hammer Testing 3.1 Description: Inspection of the structure for signs of cracking or displacement that may indicate deterioration. Dimensional survey of structural elements. Tapping surface of structure with a hammer to identify hollow sounding areas which indicate de- lamination. 3.2 Equipment Required: Access equipment required for a close inspection. Crack gauge. Surveyor s tape measure. Optical level. Hammer. 3.3 When Appropriate: Should be carried out for the assessment of all masonry arch structures, before considering whether any other NDT is required. 3.4 Form of Output: Drawings or sketches. 3.5 Advantages: May be undertaken as part of a standard inspection programme. Quick. 3.6 Disadvantages: May not identify hidden faults or features. Does not give depth of delamination. AN.2.2-6

39 Coring 3.7 Description: Holes cored into structure and samples removed to identify materials used for construction and their condition. 3.8 Equipment Required: Coring rig, pump or standpipe, generator or compressor. 3.9 When Appropriate: If accurate localised information is required. It is a good calibration tool Form of Output: Coring log Advantages: Gives internal composition and direct determination of material properties Disadvantages: Invasive. May scar the structure. Trial Pits 3.13 Description: Holes excavated into the fill above the structure to identify materials used for construction and their condition Equipment Required: Shovel. Mechanical excavator When Appropriate: If accurate localised information is required. It is a good calibration tool Form of Output: Excavation log Advantages: Gives internal composition and direct determination of material properties Disadvantages: Invasive, may be disruptive. Sonic Transmission 3.19 Description: The impact of a force hammer on one side of the structure transmits a stress wave through the structure, which is received on the opposite side by an accelerometer. The distance divided by the transit time is the average wave velocity through the structure, variations in which identify relative variations in material density. These can indicate the presence of construction details, cracks or voids, as the wave must find a path around these, resulting in a longer transit time. The depth of penetration is dependent on the size of hammer used i.e. wavelength (contact time) Equipment Required: modally tuned impact hammer fitted with a force transducer; accelerometer or displacement transducer and coupling medium eg. grease; data recording and analysis system When Appropriate: To determine the presence of flaws such as cracks or voids, inhomogeneities, and poor density areas when access is available to both sides of the structure Form of Output: The velocity is plotted in a contour map format over the area of the surface surveyed. AN.2.2-7

40 3.23 Advantages: comprehensive coverage; good depth penetration Disadvantages: does not define depth of interfaces; slow; requires skill to interpret. Sonic Tomography 3.25 Description: As for Sonic Transmission except that the tests are also performed along paths which are not perpendicular to the wall so that the masonry section is crossed by a dense net of wave paths. These provide a detailed map of wave velocity across the structure so that local values of velocity can be read across the section and the extent and location of flaws such as cracks or voids identified Equipment Required: as for Sonic Transmission; software package for tomographic reconstruction When Appropriate: When the extent and location of flaws is required, and access is available to more than one side of the structure Form of Output: Velocity contour maps over cross-sections of the structure Advantages: Interpretation easier Disadvantages: data scatter can increase residual error; very slow; requires skill to interpret. Electrical Conductivity 3.31 Description: Alternating current energises a transmitter coil, which sets up a time-varying magnetic field. This induces very small currents (eddy currents) in the structure, which in turn generate a continuously changing secondary magnetic field. This, together with the primary field, induces a current that is sensed by a receiver coil, which permits the measurement of near surface average conductivity Equipment Required: conductivity meter with transmitting and receiving coils; AN.2.2-8

41 digital data logger When Appropriate: To determine the following (although it is not always possible to identify all at the same time): relative variations in moisture content in masonry; or relative variations in conductivity; presence of hidden metallic features; possibly, height of moisture capillary rise Form of Output: Contour plots of relative average conductivity at different depths over the surface surveyed Advantages: rapid; comprehensive coverage Disadvantages: shallow depth; requires skill to interpret; may be unreliable in wet conditions. Ground Penetrating Radar (GPR) 3.37 Description: High frequency electromagnetic impulses are generated and transmitted into the structure by an antenna. These are partially reflected and refracted at each change of interface where there is a change in electrical permittivity, and recorded by a receiver. The transmitting antenna may be combined in the same housing as the receiving antenna for common monostatic (reflection mode) operation. This assembly is commonly scanned across the surface to yield a radargram trace of the reflected signals. Various antenna frequencies between typically 1.5 GHz and 100 MHz are used to investigate different structure types and material conditions Equipment Required: pulse generator; antennae; data recording system; data display When Appropriate: for deep penetration at low resolutions; for higher resolution at shallow penetrations; to detect hidden features, such as thickness of arch barrel, spandrel walls and presence of haunching (backing); AN.2.2-9

42 to detect brick ring separation, voiding to rear of masonry, variations in moisture content; to locate metal in the structure Form of Output: A two-dimensional plot of changes in level of radio energy on the cross-section Advantages: effective at mapping extent of voids or separation; rapid; comprehensive coverage Disadvantages: involves collection of a large amount of information, not all of which is of engineering significance; does not generally give the dimensions of a void; high resolution only achievable at shallow depth; requires skill to interpret; should not work in wet weather; should not be used below 0 C. Infra-red Thermography 3.43 Currently under research Description: An infra-red video camera is used to measure the amount of energy radiated from surfaces to determine the surface temperature. Variations in the pattern of heat flow can be used to identify targets within a structure Equipment Required: Thermal imaging camera 3.46 When Appropriate: Used to identify condition or walls and fill, by mapping hidden structure and defects such as delamination and moisture - but only when it affects surface emission Form of Output: Colour shaded image Advantages: Very rapid, no access required, good visually Disadvantages: Indirect measurement. Involves collection of large amount of information, not all of which is of engineering significance. AN

43 4. SELECTION OF THE MOST APPROPRIATE TECHNIQUES Preliminary Investigation 4.1 Visual inspection is the appropriate and essential first step for the surveying of masonry arch bridges. It can be used to identify the geometry of the arch, the surface condition of the masonry and the location of surface defects. 4.2 Additionally hollowness of the masonry can be identified, in conjunction with visual inspection, by hammer testing or the coin-tap-test on the surface. Trial Holes and Cores 4.3 Routine investigations, used to supplement a visual inspection, generally comprise trial holes excavated on top of the bridges, through the surfacing and fill to expose the upper surface of the barrel or any backing. If the initial inspection indicates the need, or the purpose of the survey requires it, further investigation can be carried out by coring to determine the nature and condition of the internal masonry. Cores can also be used to determine the compressive strength of the masonry. 4.4 For many structures trial holes may provide all the information that is needed. However, a core or a trial hole will only provide data for the location at which it is taken. Further Investigation 4.5 Non-Destructive Testing (NDT) techniques which can be used to extend localised information are available to give a broader picture of the hidden details of masonry arch bridges. These methods can also be used for monitoring faults and for confirming the extent of repairs. 4.6 Masonry arch bridges are highly variable in their geometry and construction. Different test methods will be appropriate for different structures and not all techniques will work in all situations. A combined testing approach is useful to gain as much information as possible in order to give the greatest confidence in the interpretation. However, the balance of testing requirements depends upon many factors such as: disruption costs, the depth of the investigation required, the extent and the nature of the structure to be investigated. Thus the requirements for each structure need to be judged individually. Ground Penetrating Radar 4.7 For identification of features at relatively shallow depth, Ground Penetrating Radar (GPR) is the appropriate technique to use. GPR will provide high resolution at shallower depths and poor resolution at deeper depths. It will penetrate voids and cracks and display features behind them. It can detect metal features, although they will then mask what is behind them. GPR can provide construction detail including the thickness of the arch barrel, spandrel walls, piers etc., brick ring separation, voiding to the rear of masonry, and variations in moisture content within construction materials. AN

44 4.8 GPR surveys carried out from the road surface may require traffic management, but could provide additional information on pavement construction, presence of haunching or saddling, location of services and the nature and condition of the fill materials within the main body of the structure. It is possible to carry out traffic speed GPR surveys, which would not cause any disruption to the traffic flow across the bridge. These could save cost but would not yield as many details as surveying from within a lane closure. Electrical Conductivity 4.9 Electrical conductivity is very good at distinguishing relative variations in moisture content within materials. Electrical conductivity will only work at shallow depths, but in providing values of relative conductivity, it can be used to determine the likely nature of the material examined and whether it is moist. It can also be used to identify the presence of hidden metallic features. Electrical conductivity may not be used in wet weather. Sonic Transmission/Tomography 4.10 For greater depths of penetration, such as through fill between parallel wing walls, sonic transmission can be used to identify relative variations in material density and provide information relating to construction detail, location of voiding or poor compaction. However, this is a much slower technique as an accelerometer or geophone must be coupled to the masonry at each receiving location and a hammer must be impacted against the structure at each transmitting location. While the presence of voids or cracks may be deduced, their size and location cannot be determined. This may be overcome by the use of sonic tomography, but at the expense of even more speed, as a network of wave paths is built up across the structure to indicate the location and extent of cracks and voids. Infra-red Thermography 4.11 Infra-red Thermography is currently under research but may be useful for mapping hidden structure and defects such as delamination and moisture. Summary 4.12 NDT techniques are not definitive, require calibration (usually by coring) and need to be used in conjunction with each other and with other methods, such as trial holes, to build up a picture of the structure and its condition. The accuracy of the results obtained by the techniques is variable and very dependent on the skill of the test operator. At present, interpretation of the test results tends to be subjective and requires experience, although new developments are making NDT more quantitative It is advisable to proceed cautiously with Non- Destructive Testing and not investigate too much of a structure, until it is established that a particular technique is producing effective results. The order of testing is not important, but all techniques need to be carried out separately during different survey sessions to maximise progress rates. Relevant Third Tier Advice Notes 4.14 Details of specific techniques are included in the Lower Tier Documents of this series of Advice Notes. Sonic Transmission and Sonic Tomography are included in Advice Note 3.2, Electrical Conductivity in Advice Note 3.4 and Ground Penetrating Radar is included in Advice Note 3.5. AN

45 5. REFERENCES 5.1 The National Roads Authority Publications NRA Design Manual for Roads and Bridges NRA Eirspan Bridge Management System 5.2 Other Publications Clarke GW (1995) Assessment of Imperfect Arches Proceedings of Sixth International Conference of Structural Faults and Repairs Das, PC An examination of masonry arch assessment methods. Proceedings IABSE Symposium: Structural Preservation of the Architectural Heritage. IABSE, Rome, 1993, Das, PC (2001) The management of masonry arch bridges, Railway Engineering-2003, London, ISBN Harvey, WJ (1999) The Complex Relationship between Analysis, Testing and Assessment, Journal of the Institution of Civil Engineers, Paper 5268 Harvey, WJ, Juhàsz, Z & F Smith (2001) Structural action of arch viaducts, Railway Engineering- 2003, London, ISBN McCann, DM & Forde, MC (2001) Review of NDT Methods in the Assessment of Concrete and Masonry Structures, NDT&E International, Elsevier Science, Vol 34, 2001, Melbourne, C (1995) Arch Bridges, Thomas Telford Page J. (1993) TRL, DoT Masonry Arch Bridges, State of the Art Review AN

46 APPENDIX A DETAILS NEEDED FOR LOAD ASSESSMENT Details Normally Visible Barrel thickness at arch side face Arch barrel material at intrados Thickness and condition of mortar joints Location of displaced voussoirs Span Profile of arch barrel Rise of arch barrel Levels of road surface and arch crown needed to determine depth of fill between road surface and arch barrel Cracks in intrados, arch side faces and parapets Details Usually Hidden Dimensions of internal sections Extent and nature of any previous strengthening Location and size of buried services Internal scour Leaching of fill Nature and condition of internal masonry Presence, position and size of internal cracks Strengthening rings or saddling Thickness of arch ring under carriageway Separation of arch rings Foundation scour AN.2.2-A/1

47 Additional Aspects to be Considered Barrel thickness may vary from that visible at the arch side faces Barrel thickness may vary in the span direction under the carriageway Bricks visible at the intrados may be of higher quality than those used in the hidden arch rings Extent of ring separation is not visible Number of rings which have separated is not visible (apart from edges) Internal form: internal ribs (spandrels) cannot be seen from the intrados. There may be stone slabs spanning across the spandrels or, alternatively, transverse arches. Between the internal ribs there may be voids or they may have been in-filled. Material properties: density of barrel masonry, fill and surfacing, compressive strength of barrel masonry, mortar type (not just the visible pointing). All of these may vary widely throughout the structure. Modifications. Unless good records exist, it can be difficult to ascertain the extent of any previous widening or strengthening. A particular problem can be lack of bond at the joint between an original arch and a new arch constructed to widen the bridge. Services may include abandoned ducts not shown on statutory undertakers searches. Notching of the barrel to fit a duct across a bridge can weaken the arch locally. The level, condition and type of backing can have a significant effect on the strength of an arch. For multi-span arches it is necessary to determine: - if the piers are solid or of shell construction; - if the pier facing brickwork is cladding to lower quality masonry; - nature of any fill to the piers; - presence, position and size of any cracks in piers; Hybrid structures: - depth of pier foundations. - unusual forms of construction, eg cast iron troughs built into arch barrel over central part of span (trial pits may weaken the structure). AN.2.2-A/2

48 ADVICE NOTE 2.3 TESTING AND MONITORING THE CONDITION OF CONCRETE STRUCTURES Contents Chapter 1. Background to this area of NDT Application 2. Formulating a Test Programme 3. Potential Testing Techniques 4. Selection of the Most Appropriate Techniques 5. References AN.2.3-1

49 1. BACKGROUND TO THIS AREA OF NDT APPLICATION Causes of Defects in Concrete Structures 1.1 This chapter describes typical defects which occur in concrete bridges. Defects in concrete structures can occur due to the following causes: inadequate design; construction errors, poor materials or workmanship; overloading; significant material deterioration; accident or fire damage; excessive or unforeseen movements; deliberate damage. 1.2 Several defects may be simultaneously occurring in a single element, complicating diagnosis. Original Construction Low Cover 1.3 Low cover will enable carbonation and/or contamination to reach reinforcement earlier, leading to more severe corrosion and consequent loss of bending, shear and axial strength. Honeycombing 1.4 Concrete honeycombed due to loss of grout and/ or fines through leakage or lack of vibration would have similar strength to no-fines concrete, giving loss of compressive, tensile and bond strength. Honeycombing is not usually very extensive as it is most often occurs locally to a formwork joint, and can be grouted up. However, it can be of particular concern at reinforcement laps and end anchorages. Cold Joints 1.5 Cold joints can result in a zone of inadequate compaction with similar results to honeycombing and consequent loss of compressive, tensile and bond strength. If a sloping cold joint occurs in a beam, then a weak zone in shear can result, particularly if the joint lies in the same direction as a potential shear failure plane. AN.2.3-2

50 Voiding 1.6 Voiding can occur below congested reinforcement with inadequate vibration. A similar effect to exposed reinforcement can occur with loss of bond between the reinforcement and the remainder of the element, which is particularly serious at laps and anchorages. Loss of concrete affects shear strength, and could affect compressive or bending strength. Voiding in a beam over the top of a column can lead to loss of vertical strength. If the void is contaminated with chlorides, particularly if there is standing water, the situation can be potentially very serious, as the reinforcement can corrode with substantial loss of section of both links and main steel. Sub-standard Concrete 1.7 Sub-standard concrete can result in effects more far reaching than simply reduced concrete strength, as poor quality concrete can be more permeable or absorbent than good concrete and consequently can take in significantly more contamination leading to serious corrosion. Insufficient Reinforcement 1.8 Insufficient reinforcement due either to inadequate design or to faulty construction involving omitted or misplaced reinforcement can lead to overstressing of the reinforcement and excessive cracking of the concrete, thereby encouraging contamination to reach the reinforcement, resulting in deterioration beyond the direct weakness caused by the inadequate reinforcement. Inadequately Anchored Reinforcement 1.9 Inadequately anchored reinforcement can lead to anchorage bond failure, splitting of the concrete through the plane of the reinforcement, and concrete cracking thereby encouraging contamination to reach the reinforcement, leading to deterioration beyond the direct weakness caused by the inadequate anchorage. Displaced Void Formers 1.10 Displacement of void formers in the casting of a member due to inadequate fixity resulting in buoyant or other movement of the void formers can lead to weakness of the member as a result of the concrete not being present in the right place, with a potential for concrete crushing and cracking. Plastic Settlement, Shrinkage and Thermal Cracking 1.11 The following types of cracking are described in the order in which they might develop in the life of the structure. Plastic Settlement Cracks 1.12 Plastic settlement cracks are caused by the attempted settlement of inadequately vibrated or compacted concrete in its fresh state which is resisted by reinforcement or formwork. In slabs, wide cracks occur directly above the reinforcement resisting settlement, tapering down from the surface to the reinforcement. These may be associated with a relatively deep layer of laitance. Cores may reveal small voids beneath the reinforcement. In columns or walls, the cracks are horizontal. Such cracks are likely to occur 10 minutes to 3 hours after placing of the concrete. AN.2.3-3

51 Plastic Shrinkage Cracks 1.13 Plastic shrinkage cracks are caused by the shrinkage of concrete and consequent cracking during the very early stages of curing. Generally occurring in slabs, they appear within 30 minutes to 6 hours of placing the concrete. Caused by curing and mix design, the nature of the cracks is influenced by reinforcement and the slab restraint. Cracks are narrow, but pass through the slab. Crazing 1.14 Crazing is the name given to a network of narrow closely spaced fine cracks which may form on the surface of concrete 1 to7 days after placing. Such cracks are rarely more than 3mm deep, and of little durability significance. They are not progressive. Early Thermal Cracking 1.15 Early thermal cracks are caused by strains induced by the heat of hydration during curing of concrete, such strains taking place before the concrete has gained sufficient strength to resist cracking. They show at between 1 day and 3 weeks, but are not progressive. They typically occur in abutments, retaining walls, wing walls and piers, and are widely spaced, and are often tapered. The cracks can be serious and penetrate deeply, and may thus cause water seepage though the member. Yielding of the reinforcement may occur where insufficient steel area and distribution is provided. Corrosion may be initiated where reinforcement is exposed where it crosses the crack. Drying Shrinkage Cracks 1.16 Drying shrinkage cracking is most common in thin slabs and walls, appearing several weeks or months after placing. Such cracks arise from a combination of the mix design and curing regime, and are influenced by the restraint of reinforcement. They may be fairly widely spaced, and parallel to reinforcement. They can lead to reinforcement corrosion. Where repairs have been carried out with shrinkage compensated concrete, after a year or so the on-going shrinkage can overcome the compensating expansion, and small cracks form. Progressive Thermal Cracking 1.17 Progressive thermal cracks are caused by strains induced by thermal movement during the life of the structure. Although they are similar in nature to early thermal cracking, to the extent that they are widely spaced, they are likely to be progressive as the cause is on-going. Caused by insufficient quantity or spacing of reinforcement, and structure geometry, they are most likely in exposed walls, parapets and slabs. Chemical Attack Carbonation Induced Reinforcement Corrosion 1.18 Carbonation is the reduction in the alkalinity of concrete due to attack from atmospheric carbon dioxide. The rate of penetration of carbonation will depend in particular upon the concrete gas permeability and consequent depth and moisture content. When carbonation reaches the depth of reinforcement, the passivity provided by the concrete can breakdown when the alkalinity falls below about 11, initiating corrosion. Carbonation induced reinforcement corrosion is therefore more common with reduced cover and/or poor quality concrete, and will affect links before main reinforcement. The effect is likely to be general over all exposed areas of an element. AN.2.3-4

52 1.19 Corrosion of the links can lead to loss of shear strength. If column binders fail due to corrosion there is a risk of vertical reinforcement buckling and consequent loss of column strength. If bursting reinforcement corrodes then splitting can occur under concentrated loading. Corrosion of holding down bolts can lead to loss of bearing fixity. Corrosion is likely to lead to cracking along the line of reinforcing bars, delamination, or spalling. Failure cracking can also occur at right angles to the line of reinforcement in cases of severe corrosion leading to loss of strength and yielding of the remaining bars in bending, shear or bursting There will frequently be visual evidence of deterioration due to reinforcement corrosion. However, testing may be useful to demonstrate reinforcement location, carbonation penetration, and to confirm the cause of visible cracking due to corrosion. Delamination of the concrete, loss of section, and bond are associated defects which may require evaluation. Chloride Induced Reinforcement Corrosion 1.21 The passivity provided by alkaline concrete is also removed by chloride ions, which may arise from contamination by de-icing salts, a marine environment, or when used within the mix as an accelerator. Road salts may spray directly over parapets and piers in particular, and leak past defective waterproofing into the bridge deck. Leakage through defective deck joints may contaminate the sub-structure A difference between carbonation induced corrosion and chloride induced corrosion is that the chlorides can only be transported when dissolved within water, hence controlling the areas affected. Once contaminated, the rate of ingress again depends upon the moisture content and its ability to move within the pores and reach the vulnerable reinforcement The structural effects of reinforcement corrosion remain similar whatever the cause of the corrosion. See under Carbonation Induced Reinforcement Corrosion above. However, chloride induced corrosion is likely to be more localized, and more likely to feature pitting. Pitting corrosion can result in complete failure of a bar in close proximity to sections in good condition. Chloride induced corrosion is also more likely to occur without external visual evidence, and in extreme cases can corrode deeply embedded bars and prestressing tendons with limited oxygen supply. Strain hardening following yield may occur at local pits while the average strain along the bar remains small. Loss of fatigue strength occurs in sections with a high proportion of variable loading Preliminary testing is principally aimed at measurement of contamination levels at incremental depths, together with parameters such as half-cell potential and resistivity that indicate the risk of corrosion. Linear polarisation resistance will indicate the rate of active corrosion The location of reinforcement relative to contaminated concrete is also important. Delamination of the concrete, loss of section, and bond are associated defects which may require evaluation. Sea water has also been known to soften the cover of concrete piers in the tidal zone, resulting in a consistency of soft putty. Cracking due to Reinforcement Corrosion 1.26 Cracking as a result of corrosion occurs along the line of the reinforcing bars, and at the edges of members where delamination of the concrete on the face reaches the edge. Failure cracking can also occur at right angles to the line of reinforcement in cases of severe corrosion leading to loss of strength and yielding of the remaining bars in bending, shear or bursting. AN.2.3-5

53 Delamination 1.27 Delamination occurs as a result of corrosion developing cracks parallel to the face of the concrete at the level of the reinforcement. Delamination can eventually occur in association with pitting corrosion but often only after the pitting corrosion is well developed. Delamination primarily effects bond strength but where it is in the compression zone it can also affect bending or compressive strength. The bond strength is particularly susceptible at laps and anchorage zones. The effect is mitigated with multiple layers of reinforcement and enhanced where there are no links to bond reinforcement to the core concrete. Column strength can be lost if binders are lapped in delaminated zones. Spalling 1.28 Spalling occurs where the cover is blown off as a result of corrosion. This can happen locally to a single bar, or as a development of delamination whereby the cover is completely removed. Corner spalls can be serious if the hooked or bobbed end of a reinforcing bar is exposed reducing the effective anchorage. Halving and Hinge Joints 1.29 Halving joints in bridge decks form stepped joints permitting rotation and possibly some longitudinal movement and sometimes enabling precast beams to be quickly lifted into place. Hinge joints in bridge decks permit rotation and carry shear and longitudinal forces through diagonal crossing bars and dowel bars at mid-depth of the slab. These joints can form a path for road salts to contaminate the concrete and lead to reinforcement corrosion, which is a particular problem as the joints rely on reinforcement alone at a discrete point for their strength. Usually it is not possible to inspect the joint or to gain access for repair and these joints are now discouraged for new construction. The use of NDT techniques is therefore especially important in helping to determine the condition of these types of joint. External Sulphate Attack 1.30 Sulphates of sodium, magnesium and calcium can react with the cement paste and result in the disintegration of the concrete. Most commonly, sulphates arise from clay sub-soils or de-icing salts. Surface deterioration and scaling of the concrete may occur, permitting ingress of other contaminants. Depending upon the composition of the cement, these chemical reactions may be accompanied by large volume changes in the concrete. Delayed Ettringite Formation (DEF) 1.31 Ettringite or calcium sulphoaluminate is formed when both sulphates and calcium aluminates in the concrete react. Rather than softening the concrete by the production of gypsum as happens with the other form of sulphate attack, the ettringite reaction, if it continues once the concrete has hardened, causes cracking similar to ASR cracking, forming macro- cracks up to 8mm wide. Delayed ettringite is caused by excessive temperatures during initial curing and can be a serious problem in precast elements cured at high temperatures. It can only be diagnosed by petrographic examination, with the core expansion test providing evidence of the potential for future expansion as with ASR. AN.2.3-6

54 Thaumasite 1.32 Thaumasite forms in concrete as a result of a reaction between calcium silicates from Portland cement and sulphates, usually from a surrounding source such as ground or seawater. The thaumasite sulphate attack eventually reduces the concrete to a pulpy mass which disintegrates, exposing the reinforcement to corrosion. Slender members are particularly vulnerable, as are those in tension where the reinforcement laps could fail in bond. Alkali Aggregate Reaction 1.33 Alkali aggregate reaction is believed to take three forms, alkali-silica, alkali-silicate and alkalicarbonate, of which alkali-silica is the most common, and arises from a reaction between the silica of certain aggregates and the alkaline pore fluid arising from high alkali cements. The gel formed by the reaction may subsequently expand causing cracking of the concrete, characterised by a star shaped pattern which may differ according to the reinforcement configuration. The cracking may be severe and may be associated with yielding of the reinforcement and, occasionally, significant weakening of the concrete. At 0.1% of alkali-silica expansion, compressive strength decreases about 12%, loss of flexural strength can be as much as 50% and the elastic modulus is reduced approximately 20%. Significant loss of bond strength can also occur. ASR can be associated with frost action, ettringite and temperature stresses in the reinforcement (which would otherwise be carried by the uncracked concrete) causing fatigue. Diagnosis is most common by petrographic examination of thin samples, with prediction of future deterioration based upon measurement of the expansion of cores in a saturated state. Treatment can vary between structural replacement, casting in additional reinforcement, and no action other than regular inspection. Chemicals 1.34 Chemicals in the air can adversely affect concrete at sites such as oil refineries and chemical works. Concrete can be significantly softened and eaten away by deposits of different chemicals. Chemicals spilt on concrete can have deleterious effects, as can those in the ground at contaminated sites such as disused or existing chemical works. Groundwater also can attack concrete if it has been contaminated by chemicals. Staining Pyrite Staining 1.35 Pyrite staining arising from inclusion of pyrite nodules close to the concrete surface. Although aesthetically displeasing, the staining should not be confused with more significant reinforcement corrosion. Lime Leaching 1.36 Lime leaching causes a build-up of white deposits on the surface of the concrete, generally along the line of cracks or honeycombed concrete. The staining is evidence of the potential for eventual corrosion if the defects extend to reinforcement, after which time the deposits may also be stained brown by corrosion products. AN.2.3-7

55 Climatic Freeze-Thaw damage 1.37 Freeze thaw damage is most common on exposed surfaces subject to wetting and frost. The network of closely spaced cracks are generally parallel to the unconstrained surface, and may appear after the first winter of exposure. The cracks will be progressive on further exposure, and usually result in lime leachate deposits on the surface. It can result in disintegration of the concrete. Voids 1.38 Voids can fill with water if there exists a path for water to seep into the voids, and the weight of water in the voids can be a serious additional loading on the structure. If the water is contaminated, serious corrosion can develop in the structure. Freezing of water in voids can result in significant overstressing of the structure. Movement Bearings 1.39 Bearings which are worn through live load fretting or other reasons, or where the sliding surfaces are prone to rusting, can develop much higher friction coefficients than those assumed at design stage, resulting in significantly higher horizontal forces being developed under temperature or traction loading, which can overstress the structure. Construction Joints 1.40 Construction joints if not detailed or constructed properly, for example with scabbled keys, can be inefficient in the transfer of shear and bending, and can allow water to percolate through, leading to corrosion. Contraction Joints 1.41 Contraction joints if not detailed or constructed properly, for example with bond breakers and sealants, can lock together thereby developing additional forces in the structure leading to overstressing and cracking, which in turn can allow water percolation and corrosion to proceed. Expansion Joints 1.42 Expansion joints if not detailed or constructed properly, for example with well aligned, bond-broken dowel bars, and suitable fillers and sealants, can lock up and/or let water through, developing additional forces in the structure leading to overstressing and cracking which can then permit water access and corrosion to develop. Movement Joints 1.43 Movement joints, if not detailed or constructed properly to permit longitudinal, shearing or rotational movement as intended, can seize up leading to overstressing, cracking and consequent deterioration and corrosion of the structure. Proprietary deck joints can leak, allowing access of contamination to sub- structures. AN.2.3-8

56 Settlement 1.44 Settlement, and particularly differential settlement, can lead to cracking of the concrete and reinforcement yielding, and in beams in association with rapidly curtailed reinforcement, can lead to more serious diagonal cracking, and associated corrosion. Walls 1.45 Reinforced concrete retaining walls under excessive earth pressure, reduced passive pressure or friction resistance or reduced bearing capacity can suffer sliding or tilting and consequent infringement of their integrity, and load carrying capacity behind the wall. Thermal Movements 1.46 Thermal movements can affect the integrity of structures, particularly if bearings have become worn and friction coefficients are higher than designed for. Piers can be pulled over and subjected to force actions for which they were not designed. Deck beams can have additional moments applied through the action of excessive bearing friction. Creep 1.47 Creep movements, if these exceed those for which the structural elements were designed, can have detrimental consequences. Deck, beams and slabs can deflect and reduce headroom and columns can shorten, sometimes having a similar effect to differential settlement. Overstressing 1.48 Overloading of bridge structures, or load exceeding the capacity of an under designed or faultily constructed structure can lead to concrete cracking or steel yielding in bending, shear or torsion; concrete crushing and steel buckling in compression; and concrete splitting and reinforcement yielding under bursting stresses below concentrated loads particularly at the top of columns. The cracking can lead to further deterioration through the resulting access for contamination Fatigue of reinforcement can lead to sudden failure if the onset of fatigue cracking through overstressing continues unaware. Hazards Impact 1.50 Impact on concrete bridges can cause cracking, loss of cover, spalling and brittle fracture, all of which can lead to corrosion of the reinforcement. Impact can also result in the dislodgement of bearings, and in the displacement and tilting of precast beams. Fire 1.51 Fire can cause a reduction in compressive strength of concrete of the order of 50% and may affect the strength of the reinforcement. Delamination and spalling of the cover concrete is typical, with consequent loss of bond strength. AN.2.3-9

57 Explosion 1.52 Explosion can result in cracking, spalling, loss of cover, bending, shear, punching shear, bearing, bursting and possibly torsion failure, and dynamically caused deterioration by overstressing without destruction. Lifting off bearings resulting in damage on re-landing can also be a problem. Ancillary Items Waterproofing 1.53 Waterproofing of bridge decks or of the backs of retaining walls if poorly detailed, applied or inadequately maintained can permit access of contaminated water leading to structural corrosion. Furniture 1.54 Parapet, safety barrier and lighting column anchorages need to be well designed and carefully installed and maintained, otherwise impact damage to these items can result in damage to the structure into which the anchorage is fixed. Corrosion of the anchorages through water ingress or electro-potential action can also affect the structure. Drainage 1.55 Drainage details need to be well designed, installed and maintained allowing for structural and temperature movements to ensure that leakage does not lead to structural corrosion. AN

58 2. FORMULATING A TEST PROGRAMME Development of the Concrete Durability Problem and Durability Testing 2.1 It is now recognised that concrete structures can be very durable providing that they have been well detailed and constructed with appropriate materials and specification. Unfortunately, the road infrastructure includes a large stock of structures built before deterioration risks were so widely understood. Those structures at risk, therefore, require carefully planned maintenance if they are to be kept in service economically and safely. Selection of appropriate remedial measures is not practicable without an understanding of conditions and deterioration mechanisms within the structure and its environment. Testing and monitoring are, therefore, of critical importance. If conditions within the material are not fully appreciated, the maintenance selected may be inappropriate. 2.2 There are many different causes of concrete deterioration, as summarised above. However, the greatest deterioration problem for the road network is reinforcement corrosion induced by de-icing salt contamination, despite the fact that widespread use of chlorides for de-icing only commenced in the 1960s. Carbonation induced corrosion is a lesser problem, arising from very low cover and/or poor concrete. In both of these processes, visual evidence only occurs many years after the deterioration mechanism commences. It is this slow deterioration that explains why the scale of the problem has only become clear over the last two decades, long after the contamination started. This fact also underlines the importance of testing at-risk structures without waiting for visual signs of distress. By the time that deterioration is obvious to all, it may be too late to adopt the most costeffective remedial measures. 2.3 As the scale of the durability problem became more apparent, so the availability of testing services and choice of methods grew. Unfortunately, engineers with a wide range of responsibilities find it difficult to keep abreast of rapid developments in such a specialist field. As a consequence, many would say that the quality of testing in the past has left something to be desired. It is to be hoped that this is now improving, as there is more advice available on the subject. In particular, for testing and monitoring concrete durability, publications include the Guide to the Testing and Monitoring of Concrete Structures by the Concrete Bridge Development Group TG2 and the NRA Eirspan Bridge Management System Manual (see References). Both of these documents provide valuable information which should be referred to in combination with NRA BA 86, which only summarises information in order to put NDT for concrete condition into context. The Investigation of Concrete Defects 2.4 The investigation of concrete defects is complicated by the fact that a number of independent causes of deterioration may combine to cause a combination of effects. Cause and effect may also be separated by a long time period. The process is, therefore, seldom one in which a single test is able to prove a single cause and effect. Thus, concrete quality, degree of compaction and reinforcement cover together provide resistance to a range of potential deterioration mechanisms in a combination of environmental influences. 2.5 The deterioration of concrete structures can have safety implications, and early diagnosis and either repair or deterioration prevention can be very cost effective. Nevertheless, it should not be presumed that all defects require testing for their diagnosis or evaluation. Testing should be proportionate to the potential consequences. Many localised defects have little aesthetic or structural consequences, or can be the subject of standard repair without testing because the cause is either obvious, small scale, or not of significance to the rest of the structure. AN

59 2.6 Tests should only be selected following a clear understanding of the role they are intended to fill. This requires some appreciation of the variety of defects which may be encountered in concrete, and their influence upon durability. Concrete deterioration may not itself be of significance, but may enable the initiation of a further deterioration mechanism that is important. Types of Testing not Covered 2.7 In covering testing and monitoring for the condition of concrete structures, this advice note does not cover construction compliance tests, load tests, or tests solely related to strength assessment. Phased Inspection and Investigation 2.8 Formulating a test programme is an important subject which deserves careful consideration if testing is to achieve its objectives. This subject is covered in full within Concrete Bridge Development Group Technical Guide 2 (CBDG TG2) chapter 2, Testing within the Structure Maintenance Process. Only a brief summary of this subject is given below. National Roads Authority s requirements for testing particular to the Eirspan Inspections are given in NRA BA35 Inspection and Repair of Concrete Road Structures. 2.9 All testing programmes should be tailored to meet a considered objective, in what may be described as a problem solving approach. It is only when the problem has been defined, and the material characteristics and changes have been identified, that the possible causes can be effectively investigated. Many different parameters can have an influence, and a structured approach is, therefore, required to avoid gathering large quantities of expensive data without homing in on the critical conclusion With regard to the role of testing within the structure maintenance process, the phases of condition monitoring, diagnosis and solution development are defined as follows: Condition monitoring: The routine process of inspecting and recording the condition of structures, including routine automated monitoring. Diagnosis: The process of deciding what is going wrong when irregularities are observed. This involves a provisional diagnosis and, when necessary, testing to confirm. Solution Development: The process of deciding what to do about detected faults, and completing the subsequent interventions. This involves solution development testing followed by repair, testing of repairs and, where appropriate, special automated monitoring. Testing within the Condition Monitoring Phase of the Inspection Process 2.11 Visual examination is the most fundamental form of condition monitoring, and forms the basis of all bridge inspections. It is only in a minority of cases that observations will lead to testing, generally where the potential consequences are significant and/or require evaluation Condition monitoring within the NRA Eirspan Inspection process takes place during Principal Inspections. Eirspan Principal Inspections include for visual inspection only. No testing is undertaken during an Eirspan Principal Inspection. Generally, testing takes place under the Special Inspections carried out in accordance with the Eirspan Bridge Management System. AN

60 Testing Within the Diagnosis Phase of the Inspection Process 2.13 Any irregularities that are observed should lead on to a provisional diagnosis. Table 2 of CBD TG2 suggests the most likely diagnosis for 25 different illustrated irregularities. If a test is required to confirm this provisional diagnosis, the most appropriate tests to use are suggested in Table 3 of CBDG TG2. Tests are not always required, particularly if the cause of the irregularity is obvious, or the consequences are not of great significance Testing within the Diagnosis Phase of the NRA Eirspan Inspection process takes place within Special Inspections. This testing may well be extended to cover that required for the solution development phase at the same time. This choice will depend upon the extent of the works and costs of access, etc. In either case, the problem solving approach in which tests are tailored to meet specific objectives remains critical. Testing Within the Solution Development Phase of the Inspection Process 2.15 Not all irregularities will require remedial action, but the choice must be made whether to: repair; monitor by automated monitoring or repeat testing; leave un-repaired and revert to Standard Inspections Testing may be required before such a decision can be taken - and in particular the specification for repair contracts should be informed by as much understanding of the extent of the problem as practicable. Decisions to monitor or to delay relevant repairs may have potentially high economic consequences, and in these circumstances, the more sophisticated NDT techniques are more likely to be justified. Table 4 of CBDG TG2 provides a summary of tests appropriate when developing a particular solution As with some Diagnosis Testing, testing within the Solution Development Phase of the NRA Eirspan Inspection process is likely to take place within a Special Inspection. The Eirspan Bridge Management System Manuals provide useful information on Special Inspection testing of particular structure types The following, Tables 2.1 and 2.2, give a selection of site and laboratory tests which can be used to investigate defects and parameters affecting durability. AN

61 Reflectometry** * Impact Echo* Galvanic Pulse Transient Analysis** * Electrical Impedence Spectroscopy** * Dynamic Testing* Radiometry*** Thermography** Electrochemical Noise*** Acoustic Emission** Radiography* Ultrasonic Pulse Velocity* Radar* Relative Humidity* Water Content* Autoclam Permeation Test* Surface Absorption* Linear Polarisation resistance* Resistivity* Half-Cell Potential* Rebound Hammer* Reinforcement Cover/ Location* Endoscope* Coring* Breakout* Delamination Survey* Carbonation* Crack Width* Visual Inspection* National Roads Authority Volume 3 Section 1 Advice Note 2.3 Test Parameter Reinforcement Corrosion Carbonation Induced Reinforcement Corrosion Chloride Induced Early Thermal Cracking Progressive Thermal Cracking Plastic Shrinkage Plastic Settlement Crazing 3 3 Drying Shrinkage Freeze Thaw Damage Alkali-Silica reaction External Sulphate Attack Delayed Ettringite Formation Thaumasite Aggregate Freeze-Thaw Pop-Outs 3 Pyrite Staining 3 1 Lime Leaching 3 Fire Damage Impact Damage Load Induced Cracking Honeycombed Concrete Voiding Key to Table: 3 = Essential /In common usage 2 = Desirable but not essential/medium Usage 1 = May be helpful in particular cases/rare Usage * = Established technique ** = Developing technique, but ready for appropriate applications *** = Technique at research stage Table 2.1 Selection of Site Tests for Concrete Condition AN

62 Density of Core* Gravimetric Measurement* Diffusion Test* X-ray Diffraction* Thermoluminescence*** Autoclam Permeation Test* Permeability* Capillary Absorption* Surface Absorption* Petrographic examination* Expansion of Cores* Alkali Content* Sulfate Content* Cement Content* Chloride Content* National Roads Authority Volume 3 Section 1 Advice Note 2.3 Test Parameter Reinforcement Corrosion Carbonation induced Reinforcement Corrosion Chloride induced Early Thermal Cracking Progressive Thermal Cracking Plastic Shrinkage Plastic Settlement Crazing 2 Drying Shrinkage 1 3 Freeze Thaw Damage Alkali-Silica reaction Sulfate attack external Sulfate attack internal (DEF) Sulfate attack internal (Thaumasite) Aggregate freeze-thaw pop-outs 3 Pyrite Staining 1 Lime leaching 2 Fire Damage 3 3 Impact Damage Load induced damage 1 Key to Table: 3 = Essential /In common usage 2 = Desirable but not essential/medium Usage 1 = May be helpful in particular cases/rare Usage * = Established technique ** = Developing technique, but ready for appropriate applications *** = Technique at research stage Table 2.2 Selection of Laboratory Tests for Concrete Condition AN

63 3. POTENTIAL TESTING TECHNIQUES Technique Description Comments/Value Inspection Tests and Samples Visual inspection Examination for cracks, discolouration, damage, integrity, environment, etc. Essential information to guide provisional diagnosis and testing programme. Hammer tapping Light tapping with a hammer to indicate hidden defects. Low cost test to detect and map delamination. Material sampling Site Tests Reinforcemen t cover Cores, dust or lump samples for laboratory testing (see Table 3.4). Electromagnetic meter to locate and give an indication of size of reinforcement and measure cover. Tests are essential, but number and location should be selected with care. Essential information at low cost. Does not change with time. Rebound hammer Surface hardness measured by constant impact. Provides rough indication of strength and uniformity. Carbonation Colour change of indicator on fresh surface or powder shows uncarbonated, hence protective concrete. Essential indicator of carbonation induced corrosion conditions. Half-cell potential Resistivity Linear polarisation resistance Surface absorption Autoclam permeation test Water content Relative humidity Electrical potential of embedded reinforcement indicates degree of risk of electrochemical corrosion. Low concrete resistivity indicates potentially higher rates of corrosion. Current required to produce a controlled potential shift is measured to determine current at rest potential. Measures rate at which water is absorbed by a dry surface under constant head. Multi-purpose equipment to measure water absorption, water permeability and air permeability, on same principles as individual tests (may also be used in the laboratory). Can be measured directly by oven dried samples, or by various indirect methods. Moisture level in concrete measured indirectly by Relative Humidity within voids. Several different methods. Standard test of risk, but requires proficient application and interpretation. Results vary with humidity of concrete. Useful supplementary test, quick to perform. Produces a direct indication of corrosion rates, but more expensive than half cell and resistivity with fewer practical applications. Unable to distinguish between severe localised corrosion and more extensive general corrosion. Practical difficulties arise on site, and variable moisture influences result, but absorption rates greatly influence contamination. Relatively robust equipment, but difficult to relate results to durability without wider application. Moisture has a significant influence on durability, but factors affecting accuracy of all methods are a concern, hence little used. Moisture has a significant influence on durability, and RH more closely related than water content. Minor damage and up to 7 days for readings. Table 3.1 Test Techniques AN

64 Technique Description Comments/Value Radar Ultrasonic pulse velocity Radiography Acoustic emission Electrochemical noise Thermography Radiometry Dynamic testing Electrical impedance spectroscopy Attenuation and reflection of electromagnetic waves indicates discontinuities and sub-surface properties. Velocity of an ultrasonic pulse is measured between transmitter and receiver on same, adjacent or opposite sides of member. Radiation from isotopes or x-rays projects image of internal features onto sensitive film on opposite side of concrete. Energy released by microcracking in a structure propagates as small amplitude elastic stress waves or acoustic emissions which can be detected as small displacements by transducers mounted on the surface. Monitoring of electrochemical noise by embedded probes. Map of surface temperatures by infra-red photography. Similar to radiography, but weak gamma rays are detected by Geiger or scintillation counter. Response of structure to shock and vibration measured. Embedded or surface probes measure corrosion rate. NDT method that can detect cover, delamination, reinforcement, voiding and sometimes contamination, moisture and material changes. Relatively expensive and difficult to interpret. Rapid coverage. Indicates variations in concrete quality and cracks and locates voiding. Expensive with safety risks. Can detect voiding and severe corrosion of reinforcement at otherwise inaccessible locations. Can monitor behaviour during load testing and detect structural cracking or post-tensioned wire breaks as they occur. Can detect micro-cracking initiated by rebar corrosion. Developing NDT method with good prospects, but would benefit from more case studies. Can locate deterioration as well as interpret cause. Corrosion processes such as passive film breakdown and pit initiation emit noise. Difficult to interpret and correlate with corrosion rates. Can provide indication of internal features, such as voiding, but effectiveness depends upon temperature history which often needs direct sunshine to provide sufficient contrast. Indicates near-surface density, and mapping features by tomography. Can indicate changes in service and demonstrate compliance with design assumptions. Specialist technique under development, not ready for application. Galvanostatic pulse transient analysis Impact echo Reflectometry Embedded or surface applied probes apply a pulse, and transient change in reinforcement potential is measured. Stress waves are reflected by internal flaws and external surfaces. Electrical signal input to end of prestressing steel, and return signal examined. Influenced by voids, etc. Research tool at an early stage of development. Can detect cracks, delamination and voiding. Difficult interpretation and only relevant to posttensioned construction. This technique is not recommended. Table 3.1 Test Techniques (continued) AN

65 Technique Description Comments/Value Automated Site Monitoring Temperature Strain Crack width Half-cell potential Resistivity Linear polarisation resistance Galvanic corrosion Moisture Laboratory Testing of Samples Chloride content Cement content Sulfate content Alkali content Expansion of cores Petrographic examination Surface absorption Capillary absorption Permeability Autoclam permeation test Thermoluminescence These parameters may be measured at frequent intervals and the data downloaded from site dataloggers or transmitted directly for remote analysis. Chloride content of samples measured preferably by titration, results generally compared with weight of cement. Cement content measured by dissolving in hydrochloric acid. Dust samples, lumps or cores analysed for sulphate content, results generally compared with weight of cement. Dust samples, lumps or cores analysed for alkali content by flame emission or atomic absorption. Expansion of core is measured whilst maintained in controlled damp warm conditions to accelerate expansion. Very thin section sample of concrete is examined by polarising petrological microscope. Cores tested to measures rate at which water is absorbed by a dry surface under constant head. Rate at which water is taken up by sample placed on wet filter papers is measured. Hole drilled into concrete is evacuated, and time for air to permeate to increase pressure to a specified level is measured. Equipment used on site or laboratory to measure water absorption, water permeability and air permeability, on same principles as individual tests. Dust samples are heated and light emission pattern is related to temperature history of certain minerals. Automated monitoring improves ability to understand the reason for variations in parameters by a study of their relationship. Low cost test providing valuable indication of risk of reinforcement corrosion. May be very inaccurate, 1kg sample required, and relatively expensive. Best avoided if some knowledge of specification already known. Test used in diagnosis of sulphate attack. Test used in investigation of alkali silica reaction diagnosed by petrographic examination. Used for diagnosis and prognosis of alkali silica reaction. Provides information on composition and durability of concrete, and for reliable diagnosis of many forms of deterioration. Relevant to durability as absorption rates greatly influence contamination. Oven dried laboratory samples more consistent than site samples. Sorptivity indicates rate at which potentially aggressive materials penetrate concrete. Permeability indicates rate at which potentially aggressive materials penetrate concrete. Relatively robust equipment, but difficult to relate results to durability without wider application. Test at development stage for investigating fire damage. Table 3.1 Test Techniques (continued) AN

66 Technique Description Comments/Value X-ray diffraction Diffusion test Gravimetric measurement Mineralogical composition of crystalline substances is determined from very small samples. Saturated core is immersed in chloride solution and depth of diffusion measured by incremental grinding and potentiometric titration. Degree of corrosion measured by weight of a reinforcing bar cleaned of corrosion products. Diagnosis tool used to support other techniques such as petrographic examination or chemical analysis. Measures susceptibility to contamination. Tool mainly used in research and occasionally to assess new concrete. Destructive method of measuring corrosion. Density of core Density of a core is measured. Used to support cement and alkali test results. Will vary according to moisture content. Analysis of concrete A range of tests are used to determine aggregate type and grading, cement content and type, sulphate and chloride content, etc. Tests are best selected according to specific need. Table 3.1 Test Techniques (continued) Summary of Selected NDT Techniques Visual Inspection 3.1 Defects Covered: Staining, cracking, spalling, impact damage, excessive deflection. Delamination if hammer tapping included. Carbonation if phenolphthalein test included. 3.2 Principles behind the Technique: The member in question is examined under adequate illumination. 3.3 Equipment Required: Adequate lighting. Additional optional equipment includes: crack width gauge, magnifying glass, light hammer, camera, phenolphthalein stick, ruler and tape for signing and located defects, cover meter, Demec gauge if studs in place. 3.4 Accuracy: Accuracy can range from subjective assessment to measurements of an accuracy required for the purpose. 3.5 Advantages: simple, quick and generally inexpensive, although costs can escalate depending on ease of access and equipment used; training may be tailored to suit the level and type of inspection to be carried out. 3.6 Disadvantages: Only surface evidence can be detected. Reinforcement, internal voids, cables and ducts cannot be examined. 3.7 Notes on use of Technique: A visual inspection is usually required before deciding on the use of NDT techniques. Acoustic Emission 3.8 Defects Addressed: Any behaviour of the structure where a pulse of energy is released such as AN

67 structural cracking, wire breaking, concrete micro- cracking due to rebar corrosion, or friction. 3.9 Principles Behind the Technique: Energy released by microcracking in a structure propagates as small amplitude elastic stress waves or acoustic emissions which can be detected as small displacements by transducers mounted on the surface Equipment Required: An array of Acoustic Emission Sensors depending on the extent of structure to be monitored, preferably manufactured under strict quality controls. Low frequency sensors are usually used for monitoring concrete structures. Cable, power supply, technical equipment, computers Accuracy: Systems with sensitivity to detect and locate signals with 1/10,000 th of the energy of a 0.5mm pencil lead break (the HSU-Nielsen field calibration standard per IS EN :2009) exceed that required for locating concrete microfracture. The effectiveness of source location by time arrival positioning can be confirmed using lead breaks as an artificial source Advantages: Powerful technique. By relating the emissions to bridge deflection under traffic loading, AE is able to distinguish between cracks which are propagating, cracks which are fretting and any known existing but inactive cracks. By its ability to distinguish between tensile cracking and compression micro- cracking, AE is able to determine, if required, with the use of a multi-sensor array and sophisticated software, whether a crack is principally a tensile crack, a shear crack, or a mixed crack. Location within a three dimensional zone of concrete is possible with a 3D array of sensors, while simple arrays provide a zonal, linear or planar location, achieved by measuring the time arrival of emissions at different sensors. AE monitoring increases the sensitivity and safety of load testing since emission commences at the early onset of damage. Can be used to monitor posttensioned wire breaks, which result in large signals, and to detect concrete microfracture resulting from the early stages of rebar corrosion. Crack development can be related to temperature changes or traffic loading Disadvantages: Expensive, but can be good value. Does not detect cracks unless they propagate or fret. In course of ongoing development for concrete bridges although over five years experience has been gained and a proprietary system with post-processor has been used for some years for detecting wire breaks. To quantify data, fretting emission needs first to be separated from growth emission. The range of fracture sizes that can be detected, and the range at which they can be detected, is dependent upon the sensor and frequency range. The amount of emission made by defects is related to their energy release or growth and not their size Notes on Use of Technique: Can be used for short or long term monitoring. A proprietary technique has recently been developed for detecting early stages of rebar corrosion. Ground Penetrating Radar (GPR) 3.15 Defects Addressed: Location of reinforcement, voids and cracking Principles behind the technique: electromagnetic echo sounding method in which a transducer (transmitter/receiver) is passed over a surface at a controlled speed. Short duration pulses of radio energy are transmitted, and the receiver detects reflections from material boundaries and features, such as embedded metalwork or voids. The amplitude, phase and velocity are influenced by the material type, and the continuity of the signal influences the shape of the components. The travel time of radio pulses is influenced by the layer thickness or depth to embedded features Equipment required: A range of different equipment specifications is available, each providing alternative combinations of cost and accuracy, and it is important that these are selected to be suitable for both the structure and its features under investigation Accuracy: The depth of features may be estimated to an accuracy of + or 10% when the pulse AN

68 velocity is calibrated by measurement of a section of similar concrete of known thickness. Velocity is influenced by variations in moisture and salinity Individual bars of all practical diameters can be identified at a spacing of 200mm c/c using a 1GHz transducer at depths up to 300mm in a typical damp concrete. Where depth is in range mm, the critical spacing is about 150mm for 32mm bars, increasing to about 200mm for 6mm bars. Below this spacing, depth becomes an issue and for practical covers, a critical spacing of about 100mm is appropriate although large bars may sometimes be difficult to separate Masking of features depends on bar size, cover and spacing. In general, it is to be expected to become critical for bars at 25-50mm cover at about 100mm spacing for small bars, but at 200mm spacing for large (32mm) bars (1GHz). It will become a problem at even greater spacings for lower frequency antennae Attempts have been made to develop techniques to identify second layers of rebar, but these have not really moved into practical site usage except on small localised areas Honeycombing can be identified, but the degree and lower surface are difficult to establish. Delamination can usually be detected, but surface breaking cracks are more problematical. Accuracy (crack width) is very difficult to judge since there are so many potential variables, unless the gap is unrealistically large Advantages: Safe to use in urban environments and quick to apply Disadvantages: Data may only be interpreted by a specialist, and processing off site is recommended with consequent delay before use of information. Results are influenced by structural features, which may complicate interpretation and require knowledge of the structure. Results are insensitive to the depth of voids. Less sensitive to voids projecting a small target size, (e.g. voids of small vertical height viewed horizontally) congested ducting and reinforcement also likely to pose problems Notes on Use of the Technique: Durability investigations usually concentrate upon diagnosis of the mechanism followed by sufficient testing to evaluate the extent of repair. Predominantly this will involve testing at shallow depths where well established economic methods are available. Radar will become more appropriate if it can be demonstrated that features, cracking and defects at depth can be reliably measured. Development is on-going, but insufficient proven examples are available to date. Impact Echo 3.26 Defects Addressed: Location of voids and cracks Principles behind the Technique: A stress (sound) wave is generated by a short duration impact on the structure. These low frequency stress waves are propagated through the structure and reflected by flaws and external surfaces, provided that they are of different acoustic impedence. Surface displacements caused by the arrival of reflected waves at the impact surface are recorded by a transducer. These are processed to display their amplitude and frequency, and may be saved for post processing Equipment Required: Ball bearing impactor, response transducer, analyser, display screen and power supply Accuracy: The thickness of concrete slabs may be obtained to an accuracy of 10%, (and therefore, presumably the depth to significant cracks perpendicular to the impact surface can also be measured to an accuracy of 10%) Advantages: Safe to use. Requires access from one side of element only. AN

69 3.31 Disadvantages: No information is obtained on the depth of void, only the presence of a crack or void. The target size must be sufficient when viewed from the direction of testing Notes on use of the Technique: In most situations, cracks are also visible on the surface and can be measured with greater certainty by simple methods. Radiography 3.33 Defects Addressed: Voiding in ducts and in structural members, and severe corrosion of reinforcing bars at inaccessible locations such as hinge and halving joints Principles behind the Technique: A source of x- rays or gamma rays is placed on one side of the concrete and a sensitive film on the other. After exposure, the film is developed to show variations in density. The time of exposure varies with the owner of the source and with one very expensive proprietary system the image can be produced in real-time on a screen Equipment Required: Radioactive source, film Accuracy: In the right conditions, loss of section can be seen on the sides of reinforcing bars, or an approximation of the loss of section on top or bottom of the bars can be obtained from the different shading of the image Advantages: Enables conditions in deep sections to be seen without breaking out the mid-depth Disadvantages: Stringent safety precautions are required. The more powerful sources require larger exclusion zones Notes on Use of Technique: Advance laboratory trials can demonstrate suitable exposure times before site use. Developing facilities on site will increase efficiency. Ultrasonic Transmission/Tomography 3.40 Defects Addressed: Voiding in ducts and voiding and cracking in concrete members Principles behind the Technique: Ultrasonic pulses are emitted by transducers, and received by transducers which may be placed on opposite or adjacent faces in transmission mode, or on the same face in reflective mode. The pulse velocity depends on the density and elastic properties of the material and will travel round voids thereby permitting determination of their location, and size if significant Equipment Required: Source of ulstrasonic pulses, receivers, and data storage device Accuracy: Ultrasonic transmission is well established for finding voids. Using tomography the presence of voids has been indicated with a grid spacing similar to the void size Advantages: Safe to use. In reflective mode, testing may be undertaken from one face. Some equipment may be used without a couplant Disadvantages: Slow to use. Accuracy using tomography not yet proven. Specialist expertise required Notes on use of Technique: Where possible, first use a faster technique such as Radar to locate areas of interest. AN

70 Infra-red Thermography 3.47 Defects Addressed: Location of voids and services, or any feature which affects the surface temperature Principles behind the Technique: The surface of the structure is photographed in infra-red to determine its temperature. Bridge decks can be photographed looking downwards from an access platform Equipment Required: Infra-red camera, access platform Accuracy: Approximate outlines of features are possible depending upon the weather conditions Advantages: Can be very useful for detecting changes in internal construction such as voids, and for discovering features such as services Disadvantages: Often requires direct sunshine to generate sufficient temperature variation to reveal the features Notes on use of Technique: Use the weather forecast to choose an appropriate day, and plan the survey at an appropriate time of day. AN

71 4. SELECTION OF THE MOST APPROPRIATE TECHNIQUES General 4.1 Most testing programmes are likely to include a number of different tests. In the case of concrete durability, particular care is required in selecting the most appropriate techniques because the available range is so wide. Some tests are tailored to particular forms of deterioration which may be unknown at the start of the investigation. Others are still at the development stage, or are appropriate only for large scale projects or the investigation of problems with potentially serious consequences. 4.2 Concrete structures are remarkably tolerant of many defects, which pose no significant risk to structural capacity. However, in a minority of cases urgent action is essential. For others, the risk is only in the very long term, but it is cost effective to intervene early. It is the job of the engineer to match the selected technique to these very variable circumstances. The potential maintenance implications of irregularities should therefore be considered at an early stage, so that the selected testing and its cost are appropriate. 4.3 A visual inspection of the structure should always precede specification of a testing programme, so that the particular circumstances of a structure can be taken into account. Specification of the same types and number of tests as used on another structure is rarely likely to be appropriate, as it is necessary to take account of: visual defects that require further investigation; the influence of access and test number on costs; the effectiveness of a staged approach in prevailing circumstances; the cost of tests relative to the potential significance of the irregularities; the time required for test results and their interpretation, if further testing is dependent upon results. 4.4 It is these considerations that support use of the problem solving approach outlined in paragraph 2.10 above. An efficient and effective test programme can only be devised if the engineer has a clear idea of what he wishes to achieve by that programme. That may be simply low level monitoring of an irregularity, or diagnosis of the cause of the irregularity, ranging up to a full investigation in order to devise and specify remedial measures. 4.5 Tables in the CBDG TG2 Reference take the reader through a test selection process, commencing with illustrated observed irregularities. It is not proposed to repeat the procedure within this advice note, but instead to emphasise the role that sophisticated and developing NDT techniques may play in specific circumstances. Concrete Decks of Steel/Concrete Composite Bridges 4.6 The concrete slab decks of steel/concrete composite structures are subject to similar defects and deterioration as the decks of concrete bridges, and are, therefore, covered by this Advice Note. The shear connectors carrying the longitudinal shear forces across the interface between the steel beams or girders and the concrete slab may suffer chloride induced corrosion or fatigue cracking or failure. While the edge of the interface can be inspected for separation or slippage, the shear connectors remain hidden and coring or NDT is required to determine their condition. Research has been carried AN

72 out in the use of Acoustic Emission for detecting shear stud fatigue cracking and further development of the technique has been carried out following a trial in the field. Monitoring of the behaviour of steel/concrete composite structures may involve half cell potential, resistivity, corrosion currents and concrete moisture content, and in relation to both the concrete and steel components, temperature and strain. Locating and Identifying Reinforcement and Physical Features 4.7 Locating physical features will seldom be required during an investigation of concrete condition, other than for reinforcement location, which can usually be achieved using a covermeter. Locating physical features is more likely during the testing to accompany strength assessment. In those rare cases where reinforcement or features require location and identification during a durability investigation, Radiography, Ground Penetrating Radar, Impact echo and ultrasonic tomography should be assessed to see if they are appropriate. Further details are given in Advice Notes 2.1, 3.1, 3.3, and 3.5. Locating Deterioration of Concrete 4.8 The diagnosis of deterioration mechanisms is likely to be best undertaken by combinations of observation, in-situ test and laboratory tests of samples. The greatest benefit of NDT tests is their ability to cover large areas of structure on a comparative basis. In this way, the results of other tests may be extrapolated to areas giving similar results, in order to predict the condition of large structures without the need for such extensive sampling. For example: 4.9 Ground Penetrating Radar (GPR) may be able to differentiate variations in material density, contamination and water content Infra-red Thermography may identify delamination and moisture variability. This method is effective where the variability affects the surface emission of infra-red Electrical Conductivity may identify variations in moisture content within materials. Detecting Ongoing Deterioration 4.12 Deterioration of concrete may be load induced, initiated by internal or external chemical attack, or other environmental influences. Conventional tests can identify the cause, but it can be particularly difficult to estimate the structural significance and on-going rate of deterioration. Acoustic emission appears likely to provide valuable information in this area as it further develops Acoustic emission registers what is happening to the structure at that particular moment, and in that sense is monitoring rather than testing. It can be used for short, intermittent or long term monitoring. Its use during load testing increases the sensitivity and safety of load testing, since emission commences at the early onset of damage. During normal service loading AE may be used to provide short term monitoring to identify fretting from pre-existing damage, and any on-going damage development. Typically, this would be for a period that includes all the loading conditions, such as one working day, or one week, and may need to be repeated under different seasonal conditions. Relating emission to loading helps data interpretation, and improves the understanding of the mechanisms causing deterioration. To detect events that occur less frequently, such as wire break or major structural fracture, much longer monitoring periods are required, and typically continuous monitoring is employed. It can be argued that after a few years monitoring with no wire breaks or major fractures detected, monitoring could be discontinued for a period until the structure has degraded further and monitoring becomes necessary. Sensors may be installed permanently and monitored continuously or periodically, the benefits of permanent installation, even for periodic testing, being that the costly access and sensor installation is a one- time requirement. AN

73 4.14 Acoustic emission can be used to home-in on structural problems. For example, an occasional noise may be heard from a viaduct, but the source may not be identifiable. Acoustic emission can be used to locate the source, perhaps from a particular joint or bearing, so that more detailed investigations can be carried out With experience by trained operatives, acoustic emission can be used to distinguish between emissions from different sources. For example, it can distinguish between friction across worn sliding surfaces of a spherical bearing and cracking of the bedding mortar below the bearing. It has been used to distinguish between cracks propagating in steel roller bearings and the emissions made as these bearings roll under the temperature movements of the structure Acoustic emission is a developing technique that will benefit from more field applications in order for results to be interpreted with greater confidence and reliability. Nevertheless, field applications have already been undertaken in the following areas AE can be used to register emissions from shear studs on steel/concrete composite construction. Laboratory research has been carried out to ascertain the suitability of AE for detecting fatigue cracks in shear stud welds. Strong emissions have also been detected from shear studs in the field and further development of the technique has been carried out following the coring out and inspection of the shear studs from which the emissions were registered. The welding of additional transverse stiffeners to existing steel beams in steel/concrete composite construction can cause weld shrinkage and pull the flanges away from the concrete slab, creating tension in the shear studs and potentially causing internal cracking of the concrete slab. AE could be used to detect such cracking and enable suitable welding procedures to be developed AE has been used for long term monitoring of post-tensioned bridges to detect and locate the positions of wire breaks. A proprietary system with post- processing tailored to this use has been in use for some years AE has been used for short term monitoring of concrete half joints to prioritise those exhibiting most damage continuing to occur under traffic loading AE has been used in the field for short term monitoring of concrete deck hinge joints to prioritise those exhibiting most damage continuing to occur under traffic loading. It has also been developed in the laboratory on concrete deck hinge joints where the background noise of emissions from corrosion of the rebars has been punctuated by greater emissions due to microcracking of the surrounding concrete caused by the rebar corrosion AE has been used in the field to detect micro- cracking in concrete as a result of rebar corrosion using the emissions from local micro-cracking of the concrete, caused by the stresses due to the formation of corrosion deposits. A proprietary system has recently been introduced for detecting the early stages of rebar corrosion Further details of these applications are given in Advice Note 3.6. AN

74 5. REFERENCES 5.1 National Roads Authority Publications 5.2 Design Manual for Roads and Bridges NRA BA 35 Inspection and Repair of Concrete Road Structures. NRA BA 86 Advice Notes on the Non-Destructive Testing of Road Structures. Advice Note 2.1 Assessing the Conditions in Grouted Ducts in Post Tensioned Concrete. Advice Note 3.1 Impact-Echo (I-E). Advice Note 3.3 Ultrasonic Transmission and Tomography for Post-tensioned Concrete Bridges. Advice Note 3.5 Ground Penetrating Radar (GPR). Advice Note 3.6 Acoustic Emission. Bridge Inspection Manual (Once Published). NRA Eirspan Bridge Management System Manuals 5.3 Other Publications Guide to testing and monitoring the durability of concrete structures. Concrete Bridge Development Group Technical Guide no 2. March Concrete Society Technical Note 60. Bungy, J.H., Millard, S.G. and Granthan, M.G. (2006). Testing of Concteye in Structures. 4 th Edition. Taylor and Francis AN

75 ADVICE NOTE 2.4 TESTING AND MONITORING THE CONDITION OF METAL STRUCTURES Contents Chapter 1. Background to this area of NDT Application 2. Formulating a Test Programme 3. Potential Testing Techniques 4. Selection of the Most Appropriate Techniques 5. References AN.2.4-1

76 1. BACKGROUND TO THIS AREA OF NDT APPLICATION Causes of Defects in Metallic Structures 1.1 This chapter describes typical defects, which occur in metallic bridges. Defects in metallic structures can occur due to the following causes: inadequate design; construction errors, poor materials or workmanship; overloading; significant material deterioration; accident or fire damage; excessive or unforeseen movement; deliberate damage. 1.2 A combination of defects may be present on a single element, making diagnosis difficult. Identification of Iron and Steel 1.3 Mild steel is most commonly used but high yield steels are used for structural members which have to carry higher stresses. 1.4 Wrought iron and cast iron are found in some of the older structures. 1.5 It is important to distinguish between wrought iron, cast iron and steel as they have different properties. Records, appearance or structural form provide useful indications to the type of material, but the only certain method is by chemical analysis and metallographic microscopy. 1.6 Steel received Board of Trade approval for use in bridge building in the UK in 1877, the use of cast iron in Ireland dates back as far as 1816, when the Ha penny Brigde was constructed. The period from about 1880 to 1900 was critical when cast iron, wrought iron and steel were all in use. All ferrous materials are magnetic except for austenitic stainless steel and some special- purpose alloys. Steel Members and Plates Deformation and Distortion 1.7 Distortions may be present in steel members or plates for a number of reasons. They could be due to initial distortions, residual stresses, lack of fit, initial out-of-flatness or out-of-straightness of the component before fabrication, external impact damage or buckling under compression loading. Distortions out of plane in the forms of waves, kinks or warping can considerably reduce the resistance to compressive forces. Any increase in distortion is significant and may reduce the loadcarrying capacity of the structure. AN.2.4-2

77 1.8 Deformation may be caused by: overstress due to inadequate design; excessive loading; increased stress from failure or yielding of adjacent components or from section loss due to corrosion of the member itself; poor detailing or fabrication causing lack of fit of a member or plate, resulting in undesigned stresses being induced during erection and use; thermal stresses and strains exceeding design limits due to problems with bearings or expansion joints; or site substitution of the wrong grade of steel or incorrectly designed bracing; fabrication defects due to poor methods of fabrication or exposure to temperature extremes during fabrication or subsequent repair works occurring during double dipping, or more commonly during welding or weld repairs; site substitution of the wrong grade of steel. Further information on distortion during fabrication is given in Tordoff (1985); impact or fire damage. Delamination 1.9 Delamination may be defined as separation into layers within the thickness of the steel in a direction parallel to the surface While the incidence of laminations has been reduced with modern steel making and rolling procedures, delamination may occur as a result of a plane of weakness being formed in the steel section during the manufacturing process. Laminations can particularly cause problems if they are at, or close to, a welded connection. Cracking and Fracture 1.11 Cracks are potential causes of complete fracture and the most common causes are fatigue and poor detailing practices producing high stress concentrations Fatigue begins with molecular level damage (known as pre-fatigue cracking) at fabrication flaws or at sites with high surface stress concentrations such as weld toes, irregular cut edges and flame cut edges and usually proceeds by a process of crack initiation and growth to failure. Cracks may range in width from hairline to sufficient to transmit light through the members Cracks may also be present in welds because of incorrect weld procedure specifications, poor welding practices and techniques, or the use of steel with poor weldability Cracking and fracture may occur as a result of impact damage which may occur several metres away from the impact site. Wear 1.15 Evidence of wear may be found in moving parts such as pins in trusses and joints. Steel Welded Connections 1.16 Welding is now widely used in the jointing of steel bridge components. Systems based on electric arc welding are now the preferred choice. Shop welding may employ active or inert gas shielding, or flux based techniques. The success of welding depends greatly on the skill of the operator and on the conditions in which the work is done. AN.2.4-3

78 1.17 Welds are of two main forms: fillet and butt welds. Fillet welds are often used for non-fatigue sensitive joints between plates inclined to each other. Butts require the jointed elements to be prepared to receive the weld. This entails chamfering either one or both edges to form a V-groove. A backing strip provided for welding from one side can be fatigue sensitive Modern weld techniques and quality management using NDT has improved weld reliability. However, construction processes and in-situ loading can still result in overstress. Welds are also susceptible to residual stresses caused by differential expansion and contraction during deposition, although particularly sensitive areas should have been stress relieved during manufacture. It is, therefore, still important to carefully inspect all welds Different quality steels require different welding procedures and filler metals. Generally, welding difficulties increase with the thickness of members and higher steel strengths. Missing Welds and Welds of Poor Quality a) Undercut Undercut is formed when a groove is melted into the parent material by the arc action and is not subsequently filled in by the weld metal. It is readily visible when not covered by a protective coating. For welds subject to fatigue stress, undercut at the toe of a weld is a serious fatigue defect: more so for fillet than butt welds. In critical applications, undercut also represents a localised loss of wall thickness. b) Porosity Porosity is caused by gas entrapment leading to rounded cavities in the weld metal and does not significantly impair the structural strength unless extensive or clustered, due to loss of wall thickness. If porosity is extensive, it may indicate other problems associated with the weld and could mask and prevent detection of more serious defects when using radiography, ultrasonics or other methods of NDT. c) Other defects Other weld defects, which are unlikely to be visible but which may be detected by specialist NDT methods, include: incomplete weld penetration; lack of side wall or root fusion; slag inclusions. See IS EN ISO 5817: Welding Fusion-welded joints in steel, nickel, titanium and their alloys (beam welding excluded) Quality levels for imperfections Further information on welding defects is given in Introduction to the Welding of Structural Steelwork (Pratt, 1989). AN.2.4-4

79 Surface Cracks in Welds or Adjacent Areas 1.21 Cracks visible at the surface are potentially the most serious form of weld defect since the stress concentration is likely to be highest at the surface. In welded members, cracks may originate within the weld and extend into the adjacent parent metal. Incipient cracks may escape detection at time of welding, but later extend under service loading. Visual inspection of welds is extremely limited and specialist NDT is likely to be necessary. Use of an illuminated magnifying glass will aid the examination of welds and adjacent areas for hairline cracks Welding is prone to cracking from fatigue, particularly at weld terminations and returns. If cracks are detected, it is likely that similar details within the structure will also be affected. Defects in welds increase the fatigue risk. a) Weld Metal Solidification Cracking Weld metal solidification cracking is widely known as hot cracking. It occurs during cooling and solidification of the weld. b) Heat Affected Zone Cracking The heat affected zone is affected by the heat input during welding and the cooling immediately afterwards. Within the HAZ the microstructure of the steel will have been affected; this may lead, in some conditions, to the steel becoming brittle and susceptible to cracking. c) Lamellar Tearing Lamellar tearing is caused by the presence of manganese sulphides and silicate inclusions which occur in steel making. When the billet is formed and rolled into plates these are extended into thin planar inclusions. Lamellar tearing can result when a large weld is made with the boundary of the weld running parallel to an inclusion. The tear occurs due to the considerable shrinkage stresses, which can occur across the thickness of the plate on cooling. Restraint due to the joint geometry and plate thickness can affect the level of stress. Tearing is generally completely below the surface and not visually detectable. Joint Slip and Tearing 1.23 Failure of connections may also occur if welds are subject to overloading or impact, or if the connection has been incorrectly designed or poorly fabricated. A close visual inspection should reveal if there is any evidence of slip. This can sometimes be seen by the development of cracks in the coating film or by signs of rubbing. AN.2.4-5

80 Steel Bolted and Riveted Connections 1.24 In older steel bridges rivets were used for joining plates together, but, with a few exceptions, have been superseded by welding and friction-grip bolting since the 1950s. The simplest form of riveted joint is the lap joint where the rivets hold together two overlapping plates but a more satisfactory joint is one in which the plates butt against each other, with cover plates fixed across the joint on both sides of the plates. The riveting process results in the rivets filling the rivet holes, ensuring no play between the rivets and the sides of the holes (unlike bolts) Bolts with nuts may be used in the same way as rivets High strength friction grip (HSFG) bolts made from high strength steels rely on the friction which develops across the contact surfaces between the plates when they are clamped together by the bolts. The bolts are tightened in a controlled manner. The surfaces of the plates in contact with each other (the faying surfaces) are usually grit blasted to ensure uniform and predictable friction. Joint Slip, Open Holes and Loose Rivets, Bolts and Nuts 1.27 A close visual examination should be made to find out whether there is any evidence of slip or movement at cover plates, washers or bolt heads or of rubbing or rusting. Slippage is particularly important in bolted connections, as it may indicate a defective joint even though the bolts appear to be tight Riveted and high-strength bolted connections in shear should be checked for condition and for loose elements including section loss to the heads of rivets and incorrect or misaligned seating or engagement of nuts on threads Since virtually no riveting is done today, restoration projects sometimes use cup-headed bolts to give the impression of riveting. An alternative procedure is to use dome-head caps over ordinary bolts. The historic fraudulent practice of dummy rivets i.e. of moulded putty with a covering of paint, was sometimes used by less scrupulous erectors. Cracks, Tears and Distortion Adjacent to Holes 1.30 Fatigue can cause both cracking and fracture of bolts, rivets and connections. Performance can be adversely affected by concentrations of stress at holes, openings and re-entrant corners. Cracks are most likely to propagate from the rivet or bolt holes Where practicable, fatigue may be confirmed by microscopic examination of a small sample. (See under Metallography in Table 3.1 below) Punching holes in steel causes work-hardening and a consequent loss of ductility. Riveted plates are vulnerable to cracking and fracture, particularly in early steels with higher levels of sulphur and phosphorus Overloading of the connection may cause plate deformation or tearing or distortion adjacent to the bolts or rivets. AN.2.4-6

81 Steel/Concrete Composite Connections 1.34 Shear connectors are provided to transfer horizontal shear (caused by loading, shrinkage and temperature differentials) between the steel beams and the concrete slab in steel/concrete composite bridges. They also anchor the slab to the beam against vertical separation. Connectors may be headed studs, bars with hoops, channels or friction grip bolts. They are protected from corrosion by the surrounding concrete, but chloride contamination of the concrete through lack or failure of waterproofing may lead to corrosion of the shear connectors particularly in the tensile zone Defects may be present in welds attaching shear connectors to the steel beams or cracks may be present or develop in these welds. Such cracks may be due to fatigue. Visual signs may not be present in the early stages of cracking. In later stages failure of shear connectors may be indicated by evidence of separation between the top flange of the steel member and the concrete. If separation has occurred the unprotected top surface of the steel or the shear connectors may be corroding. There may be evidence of separation and rust staining at the steel/concrete interface. Close visual inspection may also reveal signs of movement or rubbing between the steel and the concrete, indicating that composite action has broken down Defects in concrete parts of steel/concrete composite bridges are described in Advice Note 2.3. Corrosion Defects in Steel 1.37 Corrosion is accelerated by continuous (or even intermittent) wet conditions or by exposure to aggressive ions, such as chlorides in de-icing salts or a marine environment, and other atmospheric industrial contaminants. In these conditions, steel becomes vulnerable to both pitting and general corrosion. Pitting corrosion is a local large reduction in parent metal and can also lead to localised high stress, which may increase the risk of fatigue failure. Bi-metal contact is also a cause of corrosion. Loss of Section 1.38 Particularly vulnerable locations are areas that experience water leakage and those where water may collect, e.g. horizontal surfaces and joints Rust scale should be removed to base metal which should be measured using callipers, ultrasonic thickness meters, or other appropriate method. Deformation and Failure due to Corrosion Products 1.40 Steel corrosion products are expansive and occupy a greater volume than the parent metal. There is, therefore, a tendency to distort bolted and riveted connections, when the corrosion occurs between the faying surfaces. Bimetallic Corrosion 1.41 Bimetallic corrosion can occur where uncoated dissimilar metals are in contact in damp or wet conditions, i.e. in the presence of an electrolyte. Problems can arise when galvanised or stainless steel or non-ferrous metals are in contact with iron or carbon steel. Protective Systems for Steel 1.42 Paint systems suffer from various forms of deterioration such as cracking, flaking, chalking and peeling. Common types of failure are: blistering by solvents, water or corrosion; flaking due to smooth, loose or greasy material on the surface before painting or application of material beyond its AN.2.4-7

82 pot life; chalking due to weathering of the binder; cracking; or pin holes. Fire Damage to Steel 1.43 Steel progressively weakens with increasing temperature; the yield strength at room temperature is reduced by about 50% at 550 C, and to about 10% at 1000 C. Steel undergoes ferrite/austenite transformation above 723 C. When subjected to temperatures above 1200 C for any length of time steel becomes burnt due to the migration of low melting point constituents to grain boundaries, resulting in embrittlement Secondary effect damage can occur in bearings, movement joints and other structural members unable to accommodate the large expansions that can occur in a fire In a severe fire, unprotected steelwork will distort and will not be suitable for reuse. In a less severe fire, damage may be limited and it may be possible to retain members after checking the straightness, distortion and mechanical properties. Bolted connections often fail through shear or tensile failure or thread stripping. Any section yielding could have caused severe weakening of connections. Fatigue Damage to Steel 1.46 Fatigue crack growth under the action of repeated traffic loading is a major concern for steel road bridges. Failure may occur even though the maximum stress in any one cycle is considerably less than the fracture stress of the material. Characteristically, a fatigue fracture surface displays two distinct zones: a smooth portion indicating stages in the growth of the fatigue crack, and a rough surface, which represents the final ductile tearing or cleaving. Typically, fatigue failures do not exhibit any significant ductile necking Fatigue failure is the most common cause of cracking and fracture of steelwork structures Potential problems may exist in bridges: which have not been designed for fatigue; which have been designed to inadequate fatigue criteria; where materials and manufacturing controls may not have been adequate; where structural changes may have occurred. This may include the addition of new fixtures or repair of damage using, for example, welded cleats or brackets, flame cut holes or strengthening plates; where operational changes have occurred, such as alterations to carriageway layouts; or where there is evidence of resonance occurring in any of the structural members (resonance is when the natural frequency matches the frequency of excitation) Guidance on fatigue susceptible details can be obtained from BS 5400: Part 10: Code of Practice for Fatigue (BSI, 1980). Further Guidance on assessment for fatigue can be found in IS EN AN.2.4-8

83 Closed Steel Members Structural Hollow Sections 1.50 The internal surfaces of SHS are not normally corrosion protected as the ends are generally sealed with welded end plates, and connectors. Flaws in the sealing welds can allow the penetration of moisture and contaminants, leading to corrosion. Closed Steel Members 1.51 Closed steel members in the form of fabricated box beams or columns may either be sealed or ventilated. They should be treated as confined spaces for the purposes of access. It is not uncommon to find an accumulation of water which can overload closed members together with bird/animal excrement, and mould and fungus. All these can damage the surface protection. Additionally, frozen water will expand and can cause plate distortion or even splitting. Suspension Bridge Cables 1.52 Suspension bridge cable wires can become subject to corrosion and there is a need to investigate and monitor their condition. If water gains access within the wrapping, it can run down inside the cable and cause corrosion remote from the access point. Corrugated Steel Buried Structures 1.53 Defects in corrugated steel buried structures are generally associated with the structural or material condition. Structural condition includes alignment, cross-sectional shape and the integrity of the joints and seams whilst material condition includes thickness and soundness of protective coating, residual thickness of steel and condition of invert paving. Cast Iron Use of Cast Iron 1.54 Cast iron has only rarely been used since Fabrication elements are made by pouring the molten metal into a mould There are several types of cast iron, but that usually found in structures is known as grey, or flake graphite cast iron, from the dull grey appearance of a freshly fractured surface In addition to main structural members, cast iron was often used in handrails and balustrades, and in trusses and lattices. It was rarely used in ties, as it is brittle and relatively weak in tension Many of the defects which cast iron exhibits are similar to those of steel although it should be recognised that the homogeneity and purity of cast iron is inferior to that of present day steels. AN.2.4-9

84 Cracking and Fracture 1.58 Cracks are common defects in cast iron. They may be caused by a number of mechanisms: cooling of the metal after casting; restrained shrinkage, especially at re-entrant corners; cold spots where an earlier splash of molten iron cooled and solidified without being absorbed by further molten iron; blowhole; water accumulating in hollow members and causing cracking when it freezes; and overloading, particularly in tension Since cast iron is brittle it is liable to crack or fracture when subjected to tensile loading. Blowholes 1.60 Blowholes are common defects in cast iron. They are caused by internally trapped air reaching the surface as the iron solidifies. These may act as points of initiation for subsequent cracking. Large castings may contain hidden voids. Corrosion 1.61 During the casting process, silica in the moulding sand fuses and coats the surface of the casting, forming a barrier to oxygen. The cast surfaces of cast iron are, therefore, highly resistant to corrosion, but corrosion may still become significant because of the age of the structure. Cut surfaces, however, rust quickly in moist air Cast iron is particularly susceptible to chloride- assisted corrosion from salt (sea) water or road deicing salt. Areas within splash zones, where the cast iron is alternatively wetted and exposed to air, are particular cause for concern Cast iron can be subject to bi-metallic corrosion where it is in contact with dissimilar metals. This includes contact with galvanised (zinc-coated) steel and stainless steel (which is nickel and chrome alloy). Graphitisation 1.64 Corrosion of cast iron may occur by a process known as graphitisation. In this form of corrosion the iron is replaced by graphite with no significant change in volume but with a considerable loss of strength. Graphitisation occurs in salt or acidic water, or in ground bearing such water, and requires little oxygen. It, therefore, typically occurs below ground or water level The effect of graphitisation is to weaken the cast iron significantly, although this may not become evident until it is struck or loaded abruptly. Graphitisation can be recognised by a soft, black and blistered surface which can be easily broken away with a knife or other hard, sharp instrument. AN

85 Impact or Fire Damage 1.66 Cast iron is an inherently brittle material, with relatively poor resistance to impact. It is, therefore, likely to shatter under substantial dynamic or shock loadings. By contrast, it has a relatively high resistance to fatigue Factors influencing the tensile strength of cast iron are carbon content, rate of cooling, and size and shape of the member. Larger sections are more likely to contain flaws than smaller sections, and are therefore more susceptible to cracking and subsequent failure Cast iron does not melt at the temperatures normally encountered in a fire and temperatures up to at least 400 C do not adversely affect its basic strength. However its brittleness means that it tends to shatter when subjected to thermal shock or restraint against thermal movement. Thermal shock may typically be inflicted by applying cold water during or following a fire. Distortion 1.69 Being brittle, cast iron is likely to break rather than distort. Therefore, any significant distortions in cast iron members have usually been caused during the casting process. Wrought Iron Use of Wrought Iron 1.70 Wrought iron was rarely used following It is made by hot-rolling iron billets into elements of essentially constant cross-section, although tapered flanges may also be rolled. The manufacturing processes placed practical limitations on the size of elements so larger elements had to be built up from relatively small components, using wrought iron rivets and bolts. Closed tubes were often constructed in this way. Wrought iron was also commonly used for cables and forged links, especially in 19 th Century suspension bridges. Other applications include trusses and lattices, handrails and balustrades. Cracking 1.71 Like steel, wrought iron is a ductile material that can undergo substantial deformation before fracture, which results in characteristic necking at the fracture zone. Cracks are potential causes of complete fracture and usually occur at connections and changes in section. The most common causes are fatigue and poor detailing practices that create high stress concentrations Fatigue is likely to occur in highly stressed or reversibly stressed components. However, wrought iron has a relatively high resistance to fatigue Cracks may also be present in welds because of poor welding techniques; they are often associated with separation of the welded members Punching of holes through wrought iron members can cause local work hardening with consequent loss of ductility. This can lead to the development of cracks radiating from the holes If cracks are detected it is likely that they will be repeated in similar details within the structure Cracking and fracture may occur as a result of impact damage possibly several metres away from the impact site Deep pits, nicks or other defects may cause stress concentrations. AN

86 Wear 1.78 Evidence of wear may be found in moving parts such as pins in trusses and joints. Corrosion 1.79 Although corrosion of wrought iron is relatively slow, it may reach significant proportions because of the age of the structure. In general, the corrosion products of wrought iron cause expansion and can readily be detected Wrought iron is susceptible to chloride-assisted corrosion, typically from salt (sea) water or road deicing salt. It is especially vulnerable to this form of attack within the splash zone, where the metal is alternately wetted and exposed to air Wrought iron may also be subject to bi-metallic corrosion where it is in contact with dissimilar metals. This includes contact with galvanised (zinc-coated) steel and stainless steel (a nickel and chrome alloy). Delamination 1.82 Delamination is the separation of the material into layers within the thickness of the member in a direction parallel to the surface Delamination of wrought iron is caused by corrosion occurring along lines of slag inclusions, which, due to the rolling process, run parallel to the longitudinal axis of the element. The slag is unaffected by corrosion, so the expansive forces generated cause the rust to detach in flaky sheets. Corrosion leading to delamination occurs within the element and, therefore, deterioration may be greater than is apparent at the surface. Tapping with a hammer can provide useful qualitative information This same phenomenon causes wrought iron rods to rust into cake wedges, often with further delamination on the perimeter. Impact or Fire Damage 1.85 Wrought iron, being a ductile material, is likely to deform or distort when struck. Impact damage to a wrought iron structure is usually obvious: it can vary from scoring of the paint or metal surface to deformation of an element. In severe cases the damage may render the member incapable of carrying load Like steel, wrought iron begins to lose strength above 200 C. The effect of heating wrought iron is to anneal it, reducing its strength to that of the source metal. Therefore, a wrought iron member subjected to a fire may initially fail by its ability to sustain load being impaired Wrought iron has a high coefficient of expansion and will expand considerably during a fire In a severe fire unprotected wrought iron members will distort and will not be suitable for reuse. In a less severe fire only a few members may be affected and the straightness, distortion and mechanical properties of these can be checked. Bolted or riveted connections often fail during a fire, through shear or tensile failure or thread stripping. If there has been yielding during the fire, severe weakening of connections and sections is a possibility that will need to be considered. AN

87 Distortion 1.89 Wrought iron behaves elastically up to its yield point; further extension or deflection will cause permanent plastic deformation. Typical causes of visible distortions include: overloading causing a permanent set or buckling (elastic or inelastic); impact or fire damage; deliberate shaping or profiling; and fabrication defects, such as out-of-true rolling. Aluminium 1.90 A wide range of weldable aluminium alloys has been developed with the individual properties adjusted for a specific purpose. However, a number of characteristics are common to all these metals: high strength/density ratio, making them ideally suited to applications such as long-span structures where it is important to save weight; good general corrosion resistance, with associated low maintenance costs; readily formed and extruded into complex shapes Aluminium elements may be joined by welding, bolting or riveting. The most frequent use of aluminium is for bridge parapets, but it may occasionally be encountered used for other structural members, such as sign gantry access walkways. Corrosion 1.92 Exposure of aluminium to road salts produces a thin oxide film on the surface, which forms a protective barrier against further corrosion. However, rusting may be observed occasionally, from oxidation of iron impurities within the alloy. Some aluminium alloys are susceptible to deterioration when exposed to high concentrations of atmospheric and other environmental pollutants. These cause a change in the chemical composition, leading to corrosion or pitting. Condensation and chlorides from the surrounding environment can cause a build-up of corrosion products, including gas, within aluminium parapet posts Aluminium alloys in contact with other metals such as steel, can cause localised corrosion due to galvanic action. (In the case of aluminium and steel, aluminium acts as the anode and corrodes more.) Direct contact with concrete and other alkalis is to be avoided, and separation films such as bitumen paint have to be provided. Pitting 1.94 Localised pitting may occur due to chemical decomposition of the alloy. This generally only occurs in severely polluted industrial atmospheres. Cracking 1.95 Higher strength aluminium alloys are susceptible to stress-corrosion cracking, which can occur at stresses well below the yield stress. Such cracking gives the impression that the material is brittle, since it propagates without attendant plastic deformation. However, testing will continue to confirm that the alloy is ductile. Stress-corrosion cracking can occur as a result of residual stresses arising from manufacturing processes, which include quenching followed by machining. For further information see Shreir (1976). AN

88 Fire damage 1.96 Aluminium melts between 600 C and 660 C, and suffers a significant loss of strength above 200 C. However it has high thermal conductivity, which enables it to dissipate heat to other elements quickly, and therefore, its temperature increases relatively slowly. AN

89 2. FORMULATING A TEST PROGRAMME Maintenance Activities that may Require Support by Testing 2.1 Testing of metallic road structures may be required to support the following maintenance activities: to assess the current capacity of the structure; to assess the likely durability of the structure; to assess the feasibility and for the planning of remedial works. 2.2 In covering testing and monitoring for the condition of metallic structures, this advice note does not cover construction compliance tests, load tests, or tests solely related to strength assessment. Information Required from a Test Programme 2.3 In order to assess the current capacity, and the likely durability and to plan remedial works, the maintenance engineer will require, inter alia, the following information: the age of the structure and the dates of any modifications or repairs; the materials with which the structure has been built or modified; the properties of these materials, such as strength and ductility; the geometry of the bridge, the size of members and details of the joints; the condition of the members, joints and protection of the structure, including defects, corrosion, and distortion or damage including fatigue damage. 2.4 The bridge engineer is likely to obtain the evidence required by combining a number of different examination and test methods, each one offering particular advantages and disadvantages: visual examination including, for example, looking for corrosion, cracks, deformities, mechanical damage or indication of slip or movement at joints or at the interface between concrete deck and steel beams; testing of small samples of the materials taken from structurally safe locations to determine properties unless these are known with certainty; geometric survey to determine straightness, verticality, deformation and deflection of members, and the positions and sizes of members and details of joints; hammer tapping survey, where appropriate, to determine soundness of bolts and rivets, or corrosion leading to delamination within wrought iron members; NDT techniques to locate and/or size defects, and to determine member or paint thickness. Tests suitable for Detecting or Sizing Particular Defects 2.5 Table 2.1 is developed from Figure 3.1 in the Bridge Inspection Manual. 2.6 The advantages and disadvantages of different test techniques are summarised within Chapter 3. The set of information gathered for a given structure will inevitably be a compromise judgement between a full set of information on the one hand, and on the other hand the costs and risks of damage required to gather that information. AN

90 Table 2.1 Selection of Tests for Detection of Defects in Steel Defect or parameter Surface cracks Sub-surface cracks Internal cracks Fatigue cracks Internal voids Porosity and slag in welds Thickness of section Strain and movement Load induced cracking Material hardness Internal examination Tensile strength Ductility Material identification Weldability Surface contamination Paint film adhesion Coating thickness Coating identification Inter-coat irregularities Paint film discontinuities 3 3 = Essential 2 = Desirable 1 = May be helpful Non-Destructive Testing TEST Destructive Testing National Roads Authority Volume 3 Section 1 Hardness testing Holiday detector Paint inspection gauge Pull-off adhesion testing Cross cut/cross hatch adhesion testing Charpy or izod testing Laboratory tensile testing Metallography Chemical/sample analysis Endoscope Paint film thickness gauge Strain and deflection gauges Acoustic emission Hardness testing Radiography Ultrasonic testing Alternating current field measurement Alternating current potential drop Eddy current testing Magnetic particle inspection Dye penetrant testing Visual AN

91 Monitoring 2.7 Tests will assist in the determination of the condition of a structure at a particular time. It may be necessary or cost effective to take remedial action at that time, or it may be considered safe and economical to delay remedial action. In the latter case, it may be prudent to monitor deterioration to ensure safety is maintained and to determine the optimum time for remedial action. Monitoring may be carried out by retesting at intervals or by continuous monitoring whereby equipment is installed which records information at short intervals on a regular basis. 2.8 Continuous monitoring is a form of testing in which data logging technology is used to automatically monitor structures on an on-going basis. The most common parameters monitored are temperature, strain, and for steel/concrete composite structures, half-cell potential, resistivity, corrosion currents and concrete moisture content. Data can be collected at the structure or transmitted to a remote location. 2.9 Monitoring systems can also be designed to process data as it is being collected from the instrumentation. Hence, if the system is connected by telephone or other transmission systems, it can be designed to act as an early warning device, automatically issuing an alarm when predefined limits of the parameters are reached. This type of system can be used effectively as part of a risk management strategy Several key issues need to be addressed when considering the installation of continuous monitoring. The site equipment must be sufficiently robust to withstand elements of the weather and sited to minimise the risk of vandalism. The system will need to be maintained, the power source will need to be either battery or mains with battery back-up and the data logging capacity will need to be sufficient to store the required data between downloads. Data can be downloaded either locally, by visiting the site, or remotely, through telephone lines. Facilities will be needed for analysing the potentially large volume of data. When monitoring a structure, an accurate measure of the temperature is usually required at the same time as, for example, strain measurements. Core and surface temperature monitoring allows temperature gradients to be determined, particularly where solar radiation can induce large thermal differences, causing displacements and strains. Ambient air temperatures alone cannot fully determine differential structure temperatures There are several different methods for temperature measurement: thermocouple sensors are very versatile, resistance thermometers are very accurate and thermistors are cost-effective. AN

92 3. POTENTIAL TESTING TECHNIQUES Technique Description Comments/Value Inspection Visual inspection Survey Hammer tapping Non-Destructive Testing Dye penetrant testing Magnetic particle inspection Eddy current testing Alternating current field measurement (ACFM) Alternating current potential drop measurement (ACPD) Alternating current stress measurement (ACSM) Ultrasonic testing Radiography Examination for corrosion, cracks, deformities, damage, etc. Geometrical survey of positions and sizes of members, details of joints and trueness of members. Light tapping with a hammer to detect sound made by components. Dye highlights surface breaking cracks Iron powder indicates shallow sub-surface defects under the application of a magnetic field. Meter reads impedance change in electrical field due to sub-surface defects. A remote, uniform induced current is applied perpendicularly across the anticipated crack line, and two small sensors detect the disturbances in the magnetic field produced by the presence of a crack. An alternating current is applied perpendicularly across the anticipated crack line. The surface potential across the crack is then compared to a reference potential and the difference in potential corresponds to the crack depth. A magnetic field induced in a ferromagnetic member is affected by the stress level and sensors determine particular components of this field, providing an approximate indication of the stress. Transducer or probe converts electrical energy into ultra high frequency sound waves which are reflected by defects and recorded. X-rays or Gamma-Rays are passed through a member or weld. Voids show as darker areas on the radiograph. Essential but will not reveal fine or subsurface cracks or conditions. Essential in absence of drawings, and to check for modifications and repairs, and determine straightness, verticality, deformation and deflection of members. For use on bolts and rivets to check for looseness and on wrought iron to check for internal corrosion leading to delamination. Indicates surface cracks in members or welds not otherwise visible to the naked eye. Inexpensive. Indicates shallow (1-2mm) subsurface defects in members or welds. Magnetisation can affect subsequent welding. Sub-surface defect size can be estimated. Applicable for simple geometries. Surface-breaking defects may be detected and sized through surface protective coatings or rust layers. Useful for monitoring crack initiation in critical locations, and for monitoring the growth of known cracks. Surface-breaking defects may be detected and sized through surface protective coatings or rust layers. Useful for monitoring crack initiation in critical locations, and for monitoring the growth of known cracks. Dead load stresses in bolts, rods and strips can be determined to an accuracy of about 10% providing calibration has previously been carried out on similar members of similar material. Residual stresses in rolled joists preclude reliable stress measurements. Portable and sensitive. Can detect lamellar tearing, hydrogen cracking, solidification cracking and lack of fusion defects in members/welds. Operator skill required. Porosity in welds or castings or slag intrusions can be detected. Permanent geometrical record provided. Only defect area normal to source can be provided. Tight cracks normal to radiation can be missed. Strict safety required. Table 3.1 Test Techniques AN

93 Technique Description Comments/Value Hardness testing Acoustic emission (AE) Strain and Deflection Measurements Electrical resistance strain gauges Mechanical strain gauges Acoustic or vibrating wire gauges Inductive displacement transducers Stranded optical fibre sensors Strain gauge rosettes Blind hole drilling Diameter of imprint measured when hardened steel ball is pressed against a smooth surface with known force. Energy released by microcracking in a structure propagates as small amplitude elastic stress waves or acoustic emissions which can be detected as small displacements by transducers mounted on the surface. Wire or metal foils are bonded to the member with special adhesives. Resistance of the gauges changes with strain. Member strain produces a displacement over the gauge length which is measured by dial gauge. A sealed tube contains a taut wire plucked by an electromagnet which records the vibration. This varies with the wire tension, which is dependent on the strain in the member to which the tube is attached. An armature moving between two electrical coils enables its position to be sensed. Commercial LVDTs are available in a wide range of sizes and operating ranges from fractions of 1mm to 500mm. They are robust, linear, have high reliability, high sensitivity and an accuracy of less than ±0.5% of full scale, depending on the stroke. The resolution of LVDTs is infinitesimal and amplification of the output voltage allows detection of motions down to a few microns. The input to an LVDT can be AC. or DC, with AC. inputs more accurate, but are more complex to install and commission, so more suited to controlled environments. Light is particularly attenuated (ie reduced) when passing through microbend areas of optical fibres produced by winding three fibres around each other over sensor lengths up to 15m. An optical time domain reflectometer (by which the reflected light is recorded against time) enables the change in intensity of light pulses to determine the strain between intermediate attachment points. Cluster of strain gauges orientated in different directions. Tiny electrical resistance strain gauges can be used to measure localised strains. A small diameter hole is drilled into the member in the centre of a strain gauge rosette. This changes the existing (unknown) strain régime in a predictable way. Provides hardness number which is (only) a guide to ultimate strength. Not a positive means of identification of steel grade. Can monitor behaviour during load testing and detect fatigue damage by detecting cracks as they occur or as existing cracks fret. Increasingly well developed for steel structures, however, separation of noise from relevant emission is difficult and requires correct source location and signal analysis approach to be used. Versatile. Can be very small and read remotely. Skilled installation required. Valuable for infrequent monitoring over a long period. Manual method requiring skill. Equipment inexpensive. Gauges can be read remotely and have long term stability. They require skilled installation. Accurate method of measuring deflection. Requires signalling conditioning equipment and specialist installation. Resolution better than ±0.02mm. Care required in making connections. Cost effective monitoring of large structures. For measurement of principal strains where their direction is unknown. Enables existing or dead load strains to be determined. (This can affect member strength). Table 3.1 Test Techniques (continued) AN

94 Technique Description Comments/Value Destructive Testing Sampling Chemical Analysis Metallography Tensile testing Charpy test Izod test Information about obtaining samples from steel can be found in IS EN (general information). See also standards under Charpy and Hardness Tests below. WARNING The removal of samples for strength tests may permanently weaken the structure, particularly in fatigue of steel structures. It is essential to consider the likely value of the results in relation to possible damage to the structure and whether indirect methods of assessing strength might be more appropriate. Testing for carbon, silicon, manganese, sulphur and phosphorus to check the weldability of the steel, as a function of the carbon equivalent value as well as the impurity levels, and to provide further information on the type and associated physical properties of steel. Determination of internal structure of the material by microscopic examination of a sample with one flat surface approximately 10mm x 10mm. Tensile test on sample 200mm x 100mm (or 100mm x 50mm at extra cost) providing yield and ultimate tensile strength, modulus of elasticity and elongation at failure. Brittleness and notch ductility at a range of temperatures determined by measuring the energy required to fracture a standard U- or V-notched beam 55mm x 10mm x 10mm with a blow from a pendulum. Brittleness and notch ductility determined by measuring the energy required to fracture a standard cantilever 10mm x 10mm x 70, 98 or 126mm for one, two or three notches respectively, with a blow from a pendulum. Information about obtaining samples from welds can be found in IS EN ISO 9016 (for impact testing) and IS EN ISO 5173 (for bending tests). It is not always practicable to make sufficient tests to provide adequate data on the variability of strength, but a few tests can be useful in giving a broad indication of the quality of materials. It may be possible to remove samples from less critical areas, such as for chemical analysis. Testing carried out on drilling swarf or scrapings. Avoid contamination and ensure only material from the member under consideration is taken during sample collection. Determination of type of metal or steel or extent of fire damage from a sawn sample, and confirmation of the crack initiation and propagation mode in a failure investigation. Use of replica techniques may be used for investigating surface defects. If sample removed is by flame cutting, 10mm of heat affected zone must be removed around each sample face. See IS EN ISO Vital that sample removal does not weaken structure by reduction of cross-section, introduction of sharp corners, or embrittlement following heating. For test details see IS EN ISO and -2. More versatile than Izod test. For test details see BS 131: Part 1. Table 3.1 Test Techniques (continued) AN

95 Technique Description Comments/Value Endoscopic examination Hardness testing Close inspections of structures either by a rigid tube where viewing is by reflecting prisms, or by flexible tubes where viewing is by a fibre optic system. Illumination can be provided by glass fibres from an external source, and a camera or television monitor can be attached. Details of laboratory hardness tests are to be found in IS EN , -2 and -3 (Brinell); IS EN ISO , -2 and -3 (Vickers) and IS EN , -2, and -3 (Rockwell). Permits inspection of closed members through a small drilled hole to determine contamination, corrosion or water ingress. May also permit inspection within gaps or narrow crevices. WARNING Subsequent sealing of the holes must be carefully specified and supervised to avoid water and environmental contaminant ingress. Consideration of future examination, to determine the rate of deterioration, should also be given in the seal design. Can be used to confirm results from portable in-situ testing equipment. Table 3.1 Test Techniques (continued) Summary of NDT Techniques Visual Inspection 3.1 Defects Covered: Suitable for assessing parent metal and weld surface condition, including poor weld fit-up and misalignment; weld profile defects, including excessive or incomplete penetration, undercut, uneven or undersize fillet welds; weld spatter and stray arc strikes; surface-breaking defects such as porosity, crater cracks and gross cracking; corrosion and erosion; leaks; coating damage. 3.2 Principles Behind Technique: The element in question is visually examined under adequate illumination. 3.3 Equipment Required: Adequate lighting. Additional optional equipment includes: magnifying glass; welding gauge; fillet gauges; ruler; mirrors; boroscopes; endoscopes; fibre optics; etches, cleaners and rags; and surface replica kit to create an imprint of the surface for offsite microscopic examination. 3.4 Accuracy: Accuracy can range from subjective assessment to dimensional checks of the accuracy required for the purpose. Advantages: Simple, quick and generally inexpensive, although costs can escalate significantly according to the equipment used. Training may be tailored to suit the level and type of inspection to be carried out. 3.5 Disadvantages: Only surface-breaking defects may be detected. AN

96 3.6 Notes on the Use of the Technique: Most common non-destructive testing techniques require a visual inspection of the sample prior to testing. Visual inspection may be widened to include methods of working, such as confirming whether preheat is being correctly applied, etc. (requires clarification). Liquid (or Dye) Penetrant Testing 3.7 Defects Covered: Suitable for detecting surface- breaking defects such as cracks, laps, seams and porosity on any non-absorbent material surface. 3.8 Principles Behind Technique: A brightly coloured or fluorescent dye is allowed to enter any surface-breaking discontinuities by capillary action. Following careful cleaning away of excess dye, a suitable developer is applied which draws out the dye from the surface discontinuity. The indications formed by the dye will be wider than the underlying defect, and so will be more visible. 3.9 Equipment Required: Dye, solvent cleaner and rags, developer, illumination Accuracy: Surface defects may be accurately detected. Minimum detectable defect size is approximately 0.025mm wide Advantages: Simple, relatively quick, inexpensive and accurate. Can be used on any impermeable surface Disadvantages: Will only detect surface-breaking defects. Requires experienced operatives to identify spurious indications. Technique requires a clean surface, so all surface coatings must be removed prior to testing Notes on the Use of the Technique: Mechanical surface treatments, such as grinding, can smear the surface, closing up the defects, making them difficult to detect. Surfaces must be thoroughly cleaned before application of the dye. Adequate time must then be allowed for the dye to be drawn into the defect, and care is required when cleaning off the excess dye to ensure that the dye is not washed out of the surface discontinuities. Magnetic Particle Inspection (MPI) 3.14 Defects Covered: Suitable for detecting surface- breaking and near-surface defects such as cracks, laps, seams and porosity in ferromagnetic materials Principles Behind Technique: A magnetic field is applied to the surface to be tested, usually by means of a permanent or electromagnet. Defects on or near the surface disrupt the magnetic field, causing a flux leakage. A magnetic ink, containing fine magnetic particles such as iron filings, is applied to the surface and the magnetic particles are attracted to the area of flux leakage. This concentration of particles gives a visible indication of the defect. Fluorescent magnetic inks may be used to increase sensitivity. AN

97 3.16 Equipment Required: Solvent cleaner and rags, contrast paint, magnetising source (permanent or electromagnet, electric prods or flexible cables), magnetic ink, illumination Accuracy: Surface defects may be accurately detected. Near-surface defects may be detected, although sensitivity is reduced. Minimum detectable defect size is approximately 0.025mm wide at surface Advantages: Simple, quick, inexpensive and accurate. Surface condition is not as critical as for, say, dye penetrant testing Disadvantages: Will not detect sub-surface embedded defects. Can only be used on ferromagnetic materials, such as ferritic steel and some nickel alloys. It cannot be used on materials such as aluminium, copper or austenitic stainless steel Notes on the Use of the Technique: Defects which run parallel to the magnetic field may not be detected. Therefore, the surface to be tested should be magnetised in two directions, orientated at 90, to ensure all defects are detected. Some residual magnetisation may remain after testing. Eddy Current Testing 3.21 Defects Covered: Suitable for detecting surface- breaking and near-surface defects such as cracks in electrically conducting materials. It may also be used for detecting variations in material composition, and for measuring the thickness of non-electrically conducting coatings on electrically conducting substrates. Changes in hardness may also be detectable in some applications Principles Behind Technique: A circular coil is placed close to the sample to be tested. An alternating current (AC), typically 10Hz to 10MHz, is passed through the coil, which generates a small alternating magnetic field. This magnetic field interacts with the test sample generating small eddy currents in the surface. These currents then generate their own secondary alternating magnetic fields, which interact with the primary fields. By monitoring variations in these eddy currents, it is possible to detect near-surface defects Equipment Required: Eddy current test equipment and probes Accuracy: Surface and near-surface defects may be accurately detected, but accuracy will be affected by surface condition and surface roughness. Minimum detectable defect size is approximately 0.05mm to 0.1mm deep (assuming very good surface finish) Advantages: Quick and fairly accurate. AN

98 3.26 Disadvantages: Inspection of welds in ferritic steels can be difficult due to changes in the magnetic permeability across the weld, although special probes have been developed to lessen this effect. Will not detect sub-surface embedded defects. Can only be used on electrically conducting materials. Requires a skilled operator Notes on the Use of the Technique: Standard eddy current techniques are more generally used on non-ferritic applications, and for special applications, such as paint dry film thickness measurements. Alternating Current Field Measurement (ACFM) 3.28 Defects Covered: The equipment will detect surface-breaking defects of any width in electricallyconducting materials, including ferritic and austenitic steels and most non-ferrous materials Principles Behind Technique: The ACFM technique is a patented, non-contacting electromagnetic technique originally developed for the detection and sizing of surface-breaking fatigue cracks. A remote, uniform induced current is applied perpendicularly across the anticipated crack line, and two small sensors detect the disturbances in the magnetic field produced by the presence of a crack Equipment Required: ACFM test equipment, consisting of battery/mains powered electronic unit, a robust lap-top computer and a range of probes Accuracy: Surface-breaking cracks may be accurately detected and sized, including the depth and length of the crack. Accuracy will decrease with increased surface coating thickness. However, ACFM will be virtually unaffected by 400 micron paint coatings. Minimum detectable defect size on perfect surface is approximately 0.2mm deep or 2mm in length. Around normal fabrication welds, 10mm long x 1mm deep toe cracks or smaller can be reliably detected. Accuracy of depth sizing is generally of the order of +/- 1mm or 15% around manual welds Advantages: Quick and accurate. Does not require the removal of protective surface coatings such as paint, etc. A permanent electronic record may be obtained Disadvantages: Will not detect sub-surface embedded defects. Can only be used on electrically conducting materials. Requires a trained operator Notes on the Use of the Technique: Process may be used as an alternative to MPI and DPI, and may be used on welds through protective coatings, and will provide information on crack depths (crack face depth, not embedded depth). AN

99 Process can be used on surfaces with coatings in excess of 6mm thick, although some loss of sensitivity and accuracy will occur with thicker coatings. Alternating Current Potential Drop Measurement (ACPD) 3.35 Defects Covered: The equipment will detect and size surface-breaking defects of any width in electrically-conducting materials, including ferritic and austenitic steels and most non-ferrous metals Principles Behind Technique: An alternating current is applied perpendicularly across the anticipated crack line. A voltage probe then measures the surface potential difference. The surface potential across the crack is then compared to a reference potential. The extra path length across the crack, compared to the reference value corresponds to the crack depth Equipment Required: ACPD test equipment, consisting of battery/mains powered electronic unit, a robust computer and a range of probes Accuracy: Surface-breaking cracks may be accurately detected and sized, including the depth and length of the crack. ACPD is a technique for determining the depth of known cracks. It requires a clean surface with no rust or coatings. On a smooth surface, sizing accuracy should be within 5% Advantages: Quick and accurate. A permanent electronic record may be obtained Disadvantages: Will not detect sub-surface embedded defects. Can only be used on electrically conducting materials. Requires a trained operator. For best results the process requires a clean, smooth surface (SA 2½ minimum), due to the need for good electrical contact Notes on the Use of the Technique: ACPD is often used for sizing defects found using other techniques, such as MPI and DPI. It is also useful for monitoring crack initiation in critical locations, such as in areas of high stress or high fatigue risk, and for monitoring the growth of known cracks. Ultrasonic Testing 3.42 Defects Covered: Suitable for detecting embedded planar defects in the body of the sample, including cracks, lack of fusion defects (sidewall, inter- run, root), porosity, non-metallic inclusions and laminations in hard materials such as ferritic and non- ferritic metals and ceramics. It may also be used for wall thickness measurements and for detecting internal corrosion and pitting Principles Behind Technique: Short pulses of high frequency ultrasound (usually between 1MHz and 100MHz) are injected into the element, and the echoes received back from these pulses enable defects and wall thicknesses to be accurately detected and measured Equipment Required: Ultrasonic test equipment consisting of pulse generator, display oscilloscope AN

100 and probes Accuracy: Will accurately locate and size embedded defects Advantages: Quick and accurate Disadvantages: Requires a skilled operator. Generally limited to thicknesses of 8mm and above for weld inspection. It is necessary to obtain an echo. Some weld joint geometries do not allow a return path to enable the echo to be detected. This means that in some cases, it is not possible to fully examine all areas of the weld. This is particularly common for fillet welds Notes on the Use of the Technique: Thanks to extensive research, there are now several specific techniques available for use, depending upon the application and the structure to be examined. When manually scanning large areas, such as when carrying out lamination checks on large plates, the process can be slow and time-consuming. Radiography 3.49 Defects Covered: The equipment will detect internal defects in welds and castings in most weldable metals, including steel, cast iron, aluminium and copper. Non-metallic materials may also be examined Principles Behind Technique: A source of ionising radiation (x-rays or gamma-rays) is placed on one side of the element and a photographic plate is placed on the opposite side. The radiation is allowed to pass through the element, and is partly absorbed by the parent metal. Voids due to defects in the elements absorb less of the radiation, and will, therefore, appear darker on the film. After exposure, the photographic plate is developed to produce a two-dimensional negative image of the three-dimensional sample. This radiograph can then be viewed in a darkened room using a suitable light source Equipment Required: Radioactive isotope (gamma-ray source) and remote handling gear, or x-ray tube, light-proof cassette, photographic film, photographic development facilities, darkroom and illumination source for film assessment Accuracy: Internal defects may be detected and accurately sized. For X-rays the minimum detectable defect size in steel is approximately 0.1% of the sample thickness x 0.05mm. (max thickness 100mm). For gamma rays, the minimum detectable defect size in steel is approximately 1% of the sample thickness (max thickness 250mm) Advantages: Technique provides an accurate visual representation of the results. A permanent record is produced. Technique is particularly useful for inspecting castings. AN

101 3.54 Disadvantages: Significant safety hazards involved, requiring all personnel to be removed from the area during exposure. Cracks parallel to the film may not show up. Films are relatively expensive. Requires a skilled operator Notes on the Use of the Technique: A visual examination of the element before radiographic examination is essential to identify any defects or irregularities which could lead to an inaccurate interpretation of the radiograph. A number of radiographs of the sample must be taken from different angles to ensure that any cracks which would otherwise be parallel to the film are detected. Interpretation of radiographs can be fairly subjective, which may lead to differences of opinion. For example, differentiating between linear defects such as lack of fusion and heat affected zone cracking. Similarly, as the image is a two-dimensional representation of a three- dimensional object, surface irregularities, such as on the weld cap, can mask sub-surface defects. Hardness Testing 3.56 Defects Covered: Suitable for providing a guide to the ultimate tensile strength of the base material. May also be used to give an indication of the heat treatment condition of the base metal in some limited situations. Due to the size of indentation produced, this process is not generally suitable as a means of non- destructively testing welds. Portable, dynamic micro hardness testers may be used to measure the hardness across welds where the cap has been removed and the surface polished. However, results should be considered as guidance only Principles Behind Technique: A hardened steel ball or pyramidal diamond indenter is pressed against the smooth sample surface with a known force, and the resistance to deformation is measured as a function of the size of the indentation Equipment Required: Hardness Tester (macro or micro); for welds which have been ground flush, etches, cleaners and rags may be required to show the location of the weld and the heat affected zones Accuracy: Measures surface hardness only, so accuracy may be significantly affected by any surface effects, such as surface oxides or a decarburised layer on steel. On bright surfaces, the process is reasonably accurate for providing a guide to the tensile strength. (A decarburised layer is a thin layer of pure iron from which surface carbon, which has migrated to the surface at high temperature has been oxidised during prolonged exposure to oxygen) Advantages: Simple, relatively quick and inexpensive Disadvantages: Of limited use for non-destructively testing welds, as the heat affected zone (HAZ) is generally too narrow and the hardness gradient across the HAZ is too steep for meaningful measurement. The process requires a flat surface. Therefore, the weld cap must be ground flush to allow hardness measurements on the weld metal. AN

102 Hardness testing leaves an indentation in the surface of the sample Notes on the Use of the Technique: Hardness testing is commonly used as a destructive test technique when assessing procedure qualification and production test welds, and for monitoring base metal properties. Acoustic Emission (AE) 3.63 Defects Addressed: Any behaviour of the structure where a pulse of energy is released such as micro-cracking, overstress or friction Principles Behind the Technique: Energy released by yielding or micro-cracking in a structure propagates as small amplitude elastic stress waves or acoustic emissions which can be detected as small displacements by transducers mounted on the surface Equipment Required: An array of Acoustic Emission Sensors depending on the extent of structure to be monitored, preferably manufactured under strict quality controls such as ISO 9001 and with automated sensor test (AST) function to verify operation, operating on a single co-axial cable to supply power and carry the signal. Power supply (eg 110V, 10 amp) to technical equipment/computers Accuracy: Systems with the sensitivity to detect and locate signals with 1/10,000th of the energy of a 0.5mm pencil lead break (the HSU-Nielsen field calibration standard per IS EN : 2009) exceed that required for locating early stages of fatigue damage in steel structures. The effectiveness and accuracy of source location, which may be linear or planar, can be confirmed using lead breaks as an artificial source and adjusting the sensitivity to simulate source amplitude relative to threshold. (The source amplitude is that of the actual emissions. The threshold is the cut-off chosen to eliminate background noise. When the source amplitude is changed to that of the pencil lead break, a new threshold has to be simulated to maintain the same balance. The sensitivity is adjusted to achieve this) Advantages: Powerful technique, well established in the oil industry and increasingly used on civil structures. By relating the emissions to bridge deflection under traffic loading, AE is able to distinguish between cracks that are propagating, and any known existing but inactive cracks. May be used to home in on defects by locating their source so that other forms of NDT may be directed to the defect for sizing. Able to distinguish between friction in damaged sliding bearings and cracking of mortar bedding, and between roller bearing crack propagation and rolling friction. Defect movement can be related to temperature changes or traffic loading. AE is used following visual inspection on riveted Indian railway bridges to determine whether known or repaired cracks are still active and if so under what conditions. Rivet fretting noise is locally overcome using high frequencies, but these do not travel far, thereby precluding global monitoring. Local area monitoring by AE is also used by the Federal Highway Administration in the USA. Low access cost where aerial lift or abseiling is used to place the surface mounted, magnetically attached sensors. Large area coverage and non-intrusive nature, able to see inside structures, are significant advantages in certain situations. A derivative of Acoustic Emission, Acoustic Transmission (AT) can locate zones of severe corrosion within hollow section parapet rails Disadvantages: Cost, largely as a result of the high operator skill level and instrumentation needed to differentiate environmental noise from AE of interest, but can still be good value overall due to its large area coverage without needing entry, or damaging the paint, and low access cost where aerial lift or abseiling is used to place the sensors. Will not detect cracks unless they are propagating or fretting at the time of monitoring. On-going development for steel bridges, however, nearly ten years experience has now been obtained both in laboratory and the field on steel bridges. AN

103 Amount of emission made by active defects is not always related to their type or seriousness, this means the method is used most effectively to clear structures and areas, so that inspection effort may be focused on structures and areas with indications of active defects Notes on Use of Technique: Can be used for short or long term monitoring. Use on shear studs has been developed in the laboratory and is under development in the field. AN

104 4. SELECTION OF THE MOST APPROPRIATE TECHNIQUES 4.1 Testing programmes will inevitably include a range of tests, and should be devised with the following factors in mind: A balance must be struck between obtaining sufficient information to make a reasonable judgement on risk, and seeking so much information that the examination itself compromises durability by intrusive sampling. Tests may be effective in combination, e.g. acoustic emission for the location of cracks propagating or fretting and ultrasonics for determining their size. Tests may be interpreted in combination, e.g. a representative sample of locations revealing a particular characteristic may be examined in greater detail by a variety of more detailed tests. Testing programmes can only be provisional, and may require amendment as a result of continuing testing and interpretation. Staged testing, permitting interpretation of results between each stage, may be appropriate on larger jobs. Drilling or saw cutting and sampling can provide invaluable information on materials. However, the drilling or cutting must be carefully located, specified and supervised to avoid potentially serious damage to the structure. 4.2 It is advisable to proceed cautiously and not investigate too much of a structure until it has been established that the method is producing useful results. The effectiveness of these NDT methods may be heavily influenced by the geometry and materials of the structure under investigation. The reliability of each method will vary in particular circumstances. 4.3 Non-destructive testing techniques are not definitive, and require calibration. These Advice Notes illustrate the need to frequently evaluate the results from different tests in combination, in order to achieve meaningful interpretation. This may require a return to site after analysis of early test data. 4.4 The final objective behind a testing programme is to quantify, as far as practicable, a prediction of strength and durability. A wide range of test methods is available. The factors influencing test selection are outlined above, and the choice is complex. Nearly all structures differ in their suitability for individual methods, whether they are developing NDT or more conventional methods. In each case, the engineer must consider the prevailing circumstances, and in the light of that knowledge select the combination of tests that together builds up a picture of the structure and its condition. There will always be a need to balance cost, effectiveness, and the potential consequences that depend upon the diagnosis. 4.5 For defects addressed by various test techniques see Summary of NDT Techniques in Chapter 3 above. AN

105 Detecting Deterioration as it Occurs 4.6 Acoustic emission registers what is happening to the structure at that particular moment, and in that sense is monitoring rather than testing. It can be used for short term intermittent or long term monitoring. Experience and trials on steel bridges have shown that fatigue-related emission resulting from traffic loading is reproduced under the same loading conditions, so that monitoring for one weekday identifies the same sources (locations) as monitoring for five weekdays. For intermittent monitoring, sensors may be attached and removed for each test. However, permanent mounting and cabling to an accessible connection point provides convenient repeat testing at significantly reduced cost. Thus changes in the condition of the structure can be monitored against time, and against changes in traffic or temperature. 4.7 Acoustic emission can be used to home-in on structural problems. For example, an occasional noise may be heard from a viaduct, but the source may not be identified because the noise is being transmitted along the deck. Acoustic emission can be used to locate the source, perhaps from a particular joint or bearing, so that more detailed investigations can be carried out. Alternatively, AE can be used to locate weld fatigue cracking at locations on an orthotropic deck, so that ultrasonics can then be used to pinpoint the cracks, determine their size, and enable them to be repaired. Without AE, the resources required to examine the deck with local NDT would be very great. 4.8 With experience and using trained operatives, acoustic emission can be used to distinguish between emissions from different sources. For example, it can be used to distinguish between crack propagation and fretting of existing cracks in members or welds. Similarly, it can distinguish between friction across worn sliding surfaces of a spherical bearing and cracking of the bedding mortar below the bearing. It has been used to distinguish between cracks propagating in steel roller bearings and the emissions made as these bearings roll under the temperature movements of the structure. 4.9 AE has been used on a bascule bridge to register cracking in steel members as the bridge is lifted to provide access to vessels along the waterway below. Acoustic emission was also used to record any emissions from the low speed bearings and mechanical equipment of the bridge. AE instrumentation has also been used to identify and locate vibration and noise transients which would otherwise be difficult to find AE can be used to register emissions from shear studs on steel/concrete composite construction. Laboratory research has been carried out to ascertain the suitability of AE for detecting fatigue cracks in shear stud welds. Emissions have also been detected from shear studs in the field and further development of the technique was carried out following coring out of shear studs from which strong emissions were registered (See Advice Note 2.3 para 4.17) AE can be used to monitor suspension bridge cable wires to help detect and locate fractures resulting from corrosion For detailed applications of acoustic emission, see Advice Note A derivative of Acoustic Emission, Acoustic Transmission (AT) has been developed in the laboratory for use in determining whether hollow section parapet rails have significant internal expansive corrosion. Sonic waves are transmitted along the rail over lengths exceeding 5 metres. The signal received at the far end is calibrated to determine whether part of the rail is severely corroded. On lengths of rail so affected, ultrasonics or other local NDT can then be used to determine the locations of severely corroded areas and measure remaining thickness. AN

106 5. REFERENCES 5.1 National Roads Authority Publications 5.2 Design Manual for Roads and Bridges NRA BA 86 Advice Notes on the Non-Destructive Testing of Road Structures. Advice Note 3.6 Acoustic Emission. NRA Eirspan Bridge Management System Manual 5.3 National Standards Authority Publications IS EN IS EN IS EN 1369 Design of Steel Structures Part 1-9: Fatigue Founding Ultrasonic Examination Founding. Magnetic Particle Inspection. IS EN NDT-Acoustic Emission- Equipment characterization-equipment description. IS EN NDT-Acoustic Emission- Equipment characterization-verification of characteristics. IS EN NDT-Acoustic Emission-General Principles IS EN ISO 9016 IS EN ISO 5173 IS EN ISO IS EN ISO IS EN ISO IS EN ISO IS EN ISO IS EN ISO Destructive Tests on Welds in Metallic Materials - Impact Tests - Test Specimen Location, Notch Orientation and Examination Destructive Tests on Welds in Metallic Materials - Bend Tests Metallic Materials - Tensile Testing - Part 1: Method of Test At Room Temperature Metallic Materials - Charpy Pendulum Impact Test - Part 1: Test Method Metallic Materials - Charpy Pendulum Impact Test - Part 2: Verification of Testing Machines Non-Destructive Testing - Terminology - Part 9: Terms Used In Acoustic Emission Testing Metallic Materials - Brinell Hardness Test - Part 1: Test Method Hot Rolled Products of Structural Steels - Part 1: General Technical Delivery Conditions AN

107 IS EN ISO Hot Rolled Products of Structural Steels - Part 2: Technical Delivery Conditions For Non-Alloy Structural Steel IS EN ISO , -2, -3 Metallic Materials. Vickers Hardness Test. IS EN ISO ,- 2, -3 Metallic Materials - Rockwell Hardness Test IS EN ISO 4624 IS EN ISO 2409 Paints and Varnishes - Pull-Off Test for Adhesion Paints and Varnishes. Cross-cut Test. IS EN ISO 5817 Welding Fusion-welded joints in steel, nickel, titanium and their alloys (beam welding excluded) Quality levels for imperfections. 5.4 BSI Publications BS 131: Part 1 BS 4124 BS 5400: BS 6072 BS 7448: Part 1 The Izod Impact Test of Metals. Methods for Ultrasonic Detection of Imperfections in Steel Forgings. Steel Concrete and Composite Bridges, Part 10 Code of Practice for Fatigue. Method for Magnetic Particle Flaw Detection. Method for Determination of K Critical CTOD and Critical J Values of Metallic Materials. PD 6513 Magnetic Particle Flaw Detection. A Guide to the Principles and Practice of Applying Magnetic Particle Flaw Detection in Accordance with BS BS 7448: Part 1 Method for Determination of K Critical CTOD and Critical J Values of Metallic Materials. 5.5 Other Publications Burdekin, F M, Non-destructive Testing of Welded Structural Steel Work. Proc. Instn Civ. Engrs Structs & Bldgs, 99, Feb., Pratt, J L, Introduction to the Welding of Structural Steelwork. Steel Construction Institute, Ascot. Shreir, L L (Ed.), Corrosion, Volume 1 Metal/ Environment Reactions, Volume 2 Corrosion Control, 2 nd Edn. Newnes-Butterworths, London. Tordoff, D, Steel Bridges: The Practical Aspects of Fabrication Which Influence Design. British Constructional Steelwork Association, London. AN

108 ADVICE NOTE 3.1 IMPACT-ECHO (I-E) Contents Chapter 1. Introduction 2. Details of Technique 3. Commissioning and Specification of Impact- Echo 4. Sources of Further Information AN.3.1-1

109 1. INTRODUCTION 1.1 This Advice Note gives guidance on the use of Impact-Echo, which can be useful in detecting voids in grout in post-tensioned tendon ducts in pre-stressed concrete bridge elements. AN.3.1-2

110 2. DETAILS OF TECHNIQUE 2.1 In an impact-echo test, a transient stress pulse is introduced into the test object using a mechanical point impactor (Papoulis, 1977) - Figure 2.1. The pulse propagates through the specimen along spherical wave- fronts and along its surface. It is reflected at material boundaries. A large reflection occurs if there is a large difference in material properties and a small reflection occurs if there is a small change. The amount of reflection and transmission is given by Ghorbanpoor (1993). 2.2 The typical impact-echo set up is shown in Figure 2.1, below ( Figure 2.1 Set up for Impact-Echo Test ( After Carino 2001) 2.3 As discussed in more detail below, the impactor is typically a ball-bearing mounted on spring steel - see Figure 2.2 below: Figure 2.2 Showing the Impact-Echo Test taking place (after Colombo, Giannopoulos & Forde, 2002) 2.4 Sansalone and Streett, 1997 outline the interpretation of the impact test results. AN.3.1-3

111 2.5 If the pulse velocity through the specimen is known and the time to the arrival of a reflection from within the specimen is measured, then the distance to the target is: where: d = Depth to target (1) t = = P-wave velocity time to reflection 2.6 As the test object becomes more complex, the time history trace from the receiver becomes difficult to analyse due to multiple reflections. It is usually much more straightforward to analyse data from these tests in the frequency domain by carrying out a Fourier transform on the data. Figure Schematic diagram showing three types of response for a plate-like concrete slab containing post-tensioning ducts: a) solid plate, b) grouted duct; and c) duct containing a void (Notes: Axes: ordinate = signal amplitude [volts]; abscissa = frequency [khz]; f T = frequency associated with back wall of slab; Cp = compression wave velocity through slab) (after Sansalone and Streett, 1997) Where: λ = wavelength f = Frequency 2.7 The equations necessary to calculate the thickness of a beam (t) and the depth (d) to a specific target are given by Sansalone and Streett, 1997 in Figure 2.3. These equations assume that the time history signal has been transformed from the time domain to the frequency domain using a Fast Fourier Transform (FFT) - for easier interpretation. (Note: the original work on frequency domain testing of concrete structures relates to pile testing and the reader is referred to Davis & Dunn, 1974.) AN.3.1-4

112 2.8 The Fourier Transform is the basis of frequency analysis. The assumption is made that the time signal is made of up to an infinite number of sinusoidal components with various frequencies at different amplitudes and initial phases (Papoulis, 1977). The Fast Fourier Transform (FFT) is a highly efficient algorithm for computing the Discrete Fourier Transform and was first developed in the 1960s (Lynn & Fuerst, 1989). 2.9 The exact relationship between the highest frequency applied by the impactor (the cut-off frequency - defined by a 10dB drop in input energy) and smallest target size detectable depends upon the test material hardness (Martin & Forde, 1995). In homogeneous materials such as steel, targets oneseventh of the minimum input wavelength can be detected. However, concrete is an inhomogeneous material and so it is generally assumed that targets one half of the minimum input wavelength are the smallest detectable. Due to wave scattering from cylindrical targets, it would be expected that the detection of defects in metal ducts would be more difficult, resulting in the need for higher input frequencies. The minimum depth at which a target can be detected is assumed to be equal to half the minimum input wavelength. A reflection from a defect at this depth would have a frequency equal to the cut-off frequency and frequencies higher than this will not be measured by the testing system. This is likely to be the minimum possible depth detectable (Martin, Hardy, Usmani & Forde, 1998) An FE simulated typical output (Martin Hardy, Usmani & Forde, 1998) is shown in Figure 2.4 with features such as voids shown by a shift in the amplitude of the higher frequency components. Note Figure 2.4 is known as an A-scan or 1-D image; a B-scan would be a 2-D image as obtained using GPR. Further details of the interpretation of Figure 2.4 are contained in Martin Hardy, Usmani & Forde, Accuracy Figure Finite Element Frequency Spectra (Axes: ordinate = signal amplitude [volts]; abscissa = frequency [khz]) 2.11 The null hypothesis often used in practice is that if the impact-echo tester does not identify a target such as voiding then there is no defect. Against the background above, this null hypothesis can only be valid provided a sufficiently high frequency impulse wave is injected into the structure. It is not clear that using commercially available technology involving a ball bearing on spring steel that this objective is always fulfilled. Serious concern must exist regarding the reliability of the impact-echo technique in the absence of a load cell to measure the contact duration time of the impulse Wave scattering from cylindrical targets may increase the difficulty in detecting defects in metal ducts, resulting in the need for higher input frequencies. AN.3.1-5

113 2.13 The limitations of the effectiveness of impact- echo testing relate to: (a) (b) being able to identify a target at anything less than λ/2 to λ and the rapid attenuation of the relatively high frequency signal due to dispersion in the concrete Assuming an ultrasonic wave velocity through concrete of 4,000 ms -1 then Table 2.1 gives the minimum depth of resolution of a target. Impactor Diameter (mm) Contact Time (µsec) Maximum Usable Frequenc y (khz) Wavelengt h Min = λ (mm) Flaw Size Min Dia = λ (mm) Min Target Depth, d, min Detectable = λ/2 (mm) Min Lateral Target Size, L, Detectable = λ (mm) Max Depth of Flaw Size L dmax (mm) Table 2.1: Approximate relationships between sphere diameter, contact time, maximum useful frequency and minimum depth and lateral extent of flaws/targets (assuming Cp = 4.000m/s) (after: Note: resolution in this context relates to the depth of the shallowest identifiable target. Given the scattering due to the dispersive nature of concrete aggregate plus reinforcing bars, the actual target resolution may remain at λ/2 or even λ, as indicated in Table Note: the values of contact time in Table 2.1 may be shorter than can be achieved in the field due to increased contact time associated with crumbling of concrete (Martin and Forde, 1995). A longer contact time means a longer wavelength and poorer target resolution. Figure Scanning Impact-Echo (After Colla, 2002) AN.3.1-6

114 2.17 New research in Europe (Abraham, Le, Cote & Argoul, 2002; Colla, 2002) suggests that a B-scan (2- D image) of an I-E response can be constructed by undertaking closely spaced readings in a straight line. This B-scan would identify the response of the back- wall of the beam. It is argued that an apparently deeper back wall would indicate a voided tendon duct - since a void would mean a longer transmission path for the signal. See an example in Figure 2.5 (after Colla, 2002) There is also promising embryonic research aimed at better data analysis in Japan and Europe involving data stacking techniques (Muldoon et al 2003). Applications 2.19 The exact location of the tendons can be established from site drawings or using digital impulse radar (GPR) in the parallel configuration - see Figures 2.6, 2.7 & 2.8. The locations of the test points should be carefully recorded In Table 2.2 below, the applicability to Metal and Plastic tendon ducts is indicated respectively. I-E testing of P-T concrete beams with plastic ducts is not recommended by Sansalone & Street, 1997 but may be possible as discussed later in this Advice Note. Investigation Method Cost of Method Metal Ducts Plastic Ducts Effectiveness of Technique Visual Inspection Low No No Technique is ineffective as bridges rarely show distress before catastrophic failure. Load Test Relatively high No No Ineffective procedure and dangerous as the structure could fail before any meaningful deflection response is obtained. Stress/strain measurement Relatively high No No Generally ineffective as Cavell (1997) has shown that posttensioned bridge strain variations due to loss of pre-stressing can be similar to variations resulting from temperature gradients throughout the year. Thus, this technique is not sensitive to the defects in posttensioned bridges. Impulse radar Intermediate No Yes Effective with non-metallic liners such as in the joints of segmental bridges and in the newer post-tensioned bridges. Radar will not penetrate post-tensioned metal ducts. Impact-echo Intermediate Yes Maybe Potentially useful in identifying voiding in non-metallic and metallic post-tensioned ducts. Essential to ensure that impact frequency is sufficiently high to identify the defect. Manual drilling of tendon duct with visual inspection using endoscope. Intermediate Yes No Statistically limited and potentially dangerous if the tendons themselves are drilled. Advantage is that a direct physical observation can be made. Radiography High Yes Yes High powered radiographic techniques give good image of voiding but requires closure of the bridge and may not be used in urban areas due to risk of radiation. Ultrasonic tomography Intermediate Yes Yes Promising technique that could identify voids by producing a 2-D or 3-D image of the beam cross-section. Table 2.2 AN.3.1-7

115 Figure 2.6 Parallel Configuration: Figure 2.7 Perpendicular Configuration: 1.5 GHz antenna 900 MHz antenna Figure Illustrating Parallel and Perpendicular Configuration of Radar antennae 2.21 Impact-Echo tests should then be carried out at say 0.1m to 0.5m spacing, preferably where there are more likely to be voids in the grout in the ducts such as near the high point on a grout run, near the tendon anchor points, near the joints in precast segmental construction, or near construction joints. The location of the test points must be accurately recorded. Advantages 2.22 Impact-echo is useful for detecting voids in metallic tendon ducts - when other techniques such as radar are of no value. It may also have a role in detecting voids in plastic ducts, but radar may be the better technique. AN.3.1-8

116 Limitations 2.23 It is essential to ensure that the impact frequency, determined by the size of the ball bearing, is sufficiently high to identify the defect: (a) (b) (c) (d) (e) (f) Targets one half the minimum input wavelength are the smallest detectable. λ/2 is the minimum depth of identifiable target. Cover, must also exceed one-half of the minimum wavelength input, if the target is to be detected. Testing is hampered by the three-dimensional dispersion of the signal in concrete due to the presence of aggregate and other inhomogeneities, and performance may vary according to concrete properties. Care should be taken to ensure that the concrete surface does not crumble on ball bearing or hammer impact, otherwise the longer contact time will result in a lower frequency input signal with longer wavelength. It may be possible to reduce the masking effects of reinforcing bars near to the surface by adjusting the impact duration. Concern also exists over the reliability of the Impact-Echo technique in the absence of a load cell to measure the contact duration time of the impulse. Dense reinforcement at the anchorages will obscure the ducts at that location. Equipment 2.24 A typical Impact-Echo testing equipment set-up is shown in Figures 2.1 and 2.9. A single channel FFT analyser (or appropriate notebook PC) should be used to record the response and for performing the FFT. This analyser should measure signals at frequencies up to 100kHz. Stress waves are introduced into the test specimens using ball-bearing hammers. The response of the beam to impact should be measured using an appropriate transducer (a displacement transducer is often used). This instrument should measure surface response and should have a suitable frequency range for impactecho testing of concrete beams: D.C. to 30 khz. The response transducer must be coupled to the concrete surface using some coupling medium such as lead or grease or similar Particular care should be exercised in order that the impactor does not cause the concrete surface to crumble on impact, otherwise a longer contact time will result in a lower frequency input signal (Martin & Forde, 1995). Experience shows that multiple strikes on the same spot may help to overcome this problem The impact-echo equipment used can be bespoke or proprietary (Carino, Sansalone & Hsu, 1986; Sansalone, Lin & Carino, 1990). Details and the specification of equipment used ( and so on) should always be supplied by the test house. AN.3.1-9

117 Figure A typical impact-echo test system 2.27 The impactor and response transducer are easily portable. When working high up on a post-tensioned element, the apparatus is best operated by a team of two, with one member operating the impactor and holding the response transducer against the element, and the other member checking a suitable trace has been obtained on the laptop computer screen on the ground below. Interpretation 2.28 The null hypothesis that if the Impact-Echo tester does not identify a target such as voiding, then there is no defect, can only be valid if a sufficiently high frequency impulse wave is injected into the structure. It is not clear in using commercially available technology involving a ball bearing on sprung steel that this objective is always fulfilled. Case Study: Laboratory Experiments at the University of Edinburgh 2.29 The post-tensioned construction method allows previously unfeasible designs and new types of structures to be built. In post-tensioning the stressed reinforcement consists of steel tendons. The tendons are fed through ducts (pre-cast within the concrete structure), after the concrete has cured. They are then tensioned, compressing the structure, and anchored at each end of the member. Cement grout is then injected into the ducts, at high pressure, to form a bond between the tendons and the concrete and to fill all voids. Any voids (Figure 2.10) left within the grout may lead to corrosion of the steel tendons and ultimately collapse of the structure. AN

118 Figure Voiding of post-tensioning duct Figure Edinburgh test beam with large steel duct 2.30 The objective of this case study is to conduct laboratory experiments to verify the suitability of impact-echo to detect voids in post-tensioned concrete beams. Experimental Set-up 2.31 A number of laboratory test beams were constructed with known defects, in order to test the different investigation procedures. The beams were to offer a realistic representation of post-tensioned bridge beams, but not actually post-tensioned, and were designed to BS There are various cross-sectional shapes of post- tensioned beams, but it was decided that a rectangular cross-section would allow the greatest flexibility for testing. In addition, rectangular cross-section would not hinder data interpretation, as would a complex shape. A minimum quantity of steel was included in the design to increase durability, increase realism and for safety. The final design of the beams dictated dimensions of 400mm (Wide) x 450mm (Deep) x 2000mm (Long) with approximately 230mm to the duct (Figures 2.11, 2.8 & 2.12). AN

119 Figure Beam Details AN

120 2.33 Once the overall design of the beams was finalised, duct sizing was chosen. Of the ten beams to be constructed, five were to have steel ducting and the other five plastic The ducts were filled with different levels of grout, producing an interface that could be examined. The grouting was completed by rotating the beams vertically. Grout was pumped in under pressure, through a rubber tube, until the level of grout was 1 metre (half-way) up the beams, for the fully grouted part (Figure 2.13). Once the grout had set, the beams were realigned horizontally and filled with grout until the levels at the open end were as required. The levels chosen were: 100% grouted 100% grouted / 50% grouted Fully voided (a) (b) Figure 2.13 Beam grouting (a) grouting of large steel duct, (b) level checking using 1m ruler Experiments: 2.35 The appropriate ball bearing was chosen, by comparing the required resolution and thus the required wavelength (higher resolution = short wavelength = smaller ball bearing) and the depth of penetration needed (greater penetration = longer wavelength = larger ball bearing). The appropriate ball bearing to use on the beams was the 10mm diameter ball bearing (Martin, 1997) The velocity of the concrete was calculated by using one transducer, and the equation in Figure 2.3. The velocity is measured by impacting the surface and recording the frequency over an area of solid concrete. To simulate testing on site, the side of the beam was tested rather than the top as the top would not be accessible on site. The frequency of the rear wall (f T ) was found to be the same on all the beams tested, as all the beams where cast from the same batch of concrete. The f T was found to be 4.9 khz and thus the velocity through the concrete (C p ) was found to be 4083ms From this initial calculation it was then possible to calculate the expected position of the f void and the f steel. f void = 13 khz, f steel = 6 khz AN

121 2.38 Figure 2.14 shows the frequency response from a voided duct whilst Figure 2.15 shows the frequency response from a fully grouted duct. These results were taken at the middle of the concrete beam to reduce any end effects. It can be seen that from Figure 2.14 the initial peak (f T ) has moved forward from 4.9 (plain concrete), and there is a peak with a higher frequency, these being typical of a voided duct. In Figure 2.15 the initial f T has not moved forward and there is a pulse at 6 khz. Figure Result of an impact-echo test over an ungrouted (voided) tendon duct Notes: (a) Top half of figure is a time domain plot: ordinate = amplitude [units = + volts]; abscissa = time [units = µs]; (b) Bottom half of figure is a frequency domain plot: ordinate = amplitude [volts]; abscissa = frequency [khz] AN

122 Figure Result of an impact-echo test over a grouted tendon duct Notes: (a) Top half of figure is a time domain plot: ordinate = amplitude [units = + volts]; abscissa = time [units = µs]; (b) Bottom half of figure is a frequency domain plot: ordinate = amplitude [volts]; abscissa = frequency [khz] 2.39 From these results it is noted that the results are repeatable. The comparative difference between a voided and a grouted duct is easily identifiable. The percent success rate of clean inputted signal and usable response signals are variable depending on the operator. Unusable wave forms are easily distinguishable from good, repeatable wave forms A beam which was fully grouted except for one half - which was ½ grouted and ½ voided, was tested next (naturally the top half of this section of the duct was voided). The entire length of the beam was tested to see if the method could identify the location of the partially grouted duct. Figure 2.16 shows the impact- echo result from a typical area of this partially grouted duct. By testing the entire length it was possible to show experimentally that the ends would have an effect on the result and Figure 2.17 shows a result from near the end of the beam. AN

123 Figure Result of an impact-echo test over a half grouted tendon duct Figure Result of an impact-echo test over a grouted tendon duct at one end showing end effects AN

124 2.41 From this experiment it was possible to state that experimentally there is an end effect which needs to be taken into consideration during testing as it might bring up anomalies. Also it is possible to identify areas that are ½ voided Even though Sansalone and Streett (1997) claim that impact-echo does not work on plastic ducts, it was decided to conduct tests on plastic ducts where it is fully voided (Figure 2.18) and fully grouted (Figure 2.19). Figure Result of an impact-echo test over a fully voided plastic duct Figure Result of an impact-echo test over a fully grouted plastic duct 2.43 The results from these tests show that it may be possible to tell the difference between a fully grouted and fully voided duct even when the duct is made from plastic. Note, however, that radar may be a better test method. AN

125 3. COMMISSIONING AND SPECIFICATION OF IMPACT-ECHO Introduction 3.1 The following are specific requirements for the use of techniques for the assessment of the grouting of post-tensioned structures using impact-echo and should be used in conjunction with those given in Advice Note 1 General Guidance of this series of Advice Notes. Information Supplied to Tenderers 3.2 The tender documents submitted to testing organisations should require the following additional information to be provided: as-built drawings providing reinforcing details, duct types and sizes, profiles, locations and depth of cover; which ducts and which lengths of ducts are to be tested in detail; whether other ducts or intermediate lengths of the above ducts are to be tested at sample locations. Interval to be stated; whether the tendon ducts are metallic or plastic. 3.3 The following additional deliverables from the testing organisations should be stated: for each test location whether the techniques used have been able to detect a void in the duct; the minimum size of void that it would have been possible to detect; whether any ducts were too close to the surface for voids to be detected. Information Required of Tenderers 3.4 The Tenderers should be asked to state the following additional information: ball sizes and frequencies which will be used. How to Use the Results Additional Considerations 3.5 The following limitations should be considered. dense reinforcement may affect the accuracy of the readings, or outer ducts may mask the readings from inner ducts; vacuum testing rather than sonic techniques should be used to determine the volume of the voids in grouted ducts as a precursor to regrouting them. Guidance documents have been prepared on regrouting techniques (Ref Concrete Society Technical Report TR72 Durable bonded post-tensioned concrete structures 2010). AN

126 Impact-Echo Report 3.6 The report should include the following additional requirements: the likely impact frequency achieved and the ball bearing sizes used at each location; a discussion of the ball sizes and frequencies used and a demonstration of whether these were appropriate for locating voids at all locations investigated. AN

127 4. SOURCES OF FURTHER INFORMATION Abraham, O, Le, T-P, Cote, Ph & Argoul, P. (2002) Two enhanced complementary impact-echo approaches for the detection of voids in tendon ducts, IABMAS 02, July 2002, UPC, Barcelona, CD-Rom. ACI Technical Report 228.2R-98 (1998) Nondestructive Test Methods for Evaluation of Concrete in Structures, ACI, Farmington Hills, MI, USA, p. 62. Carino, N.J. (2001) The impact-echo method: an overview, Proc 2001 Structures Congress & Exposition, May 21-23, 2001,Washington, D.C., ASCE, Reston, Virginia, Peter C. Chang, Editor, p. Carino, N.J. & Sansalone, M. (1984) Pulse-Echo Method for Flaw Detection in Concrete. NBS Technical note U.S. Dept. of Commerce/National Bureau of standards pp34. Carino, N.J. & Sansalone, M. (1992) Detection of voids in grouted ducts using the impact-echo method, ACI Materials Journal, Vol 89, No 3, May/June, Carino, N.J., Sansalone, M. & Hsu, N.N. (1986) A Point Source-Point Receiver, Pulse-Echo Technique for Flaw Detection in Concrete. ACI Journal, March-April 1986, Vol. 83, No. 2, Cavell, D.G. (1997) Assessment of deteriorating post- tensioned concrete bridges, PhD thesis, University of Sheffield. Colla, C., Schneider, G. & Wiggenhauser, H. (2000) Automated Impact-Echo and method improvement: 2 and 3-D imaging of concrete elements. Non- Destructive Testing in Civil Engineering 2000, ed Uomoto T, Elsevier, Colla, C. (2002) Advanced impact-echo technique for NDE of post-tensioned concrete, IABMAS 02, July 2002, UPC, Barcelona, CD-Rom. Colombo, S., Giannopoulos, A. & Forde, M.C. (2002) Accuracy of radar testing of masonry arch bridges, IABMAS 02, July 2002, UPC, Barcelona, CD-Rom. Concrete Society Technical Report TR72 (1996), Durable Bonded Post-Tensioned Concrete Structures, The Concrete Society, Slough, UK, pp 64. Davis, A.G. & Dunn, C.S. (1974) From Theory to Field Experience with the Non-destructive vibration testing of piles. Proceedings of the Institution of Civil Engineers, Part 2, 1974, Vol. 57, Dec., DTP Press Notice No (1992) Published by DTP, London Ghorbanpoor, A. (1993) Evaluation of Post-tensioned Concrete Bridge Structures by the Impact- Echo Technique. U.S. Department of Transportation. Federal Highway Administration. Publication No. FHWA-RD Dec pp84. Kearey, P. & Brooks, M. (1991), An Introduction to Geophysical Exploration, Oxford. Lausch, R., Knapp, J. & Colla, C. (2001) Influence of impact source frequency on Impact-Echo data from testing of concrete structures. Proc. Struct. Faults and Repair 2001, Eng. Technics Press. AN

128 Lynn, P.A. & Fuerst, W. (1989) Introductory Digital Signal Processing with Computer Applications. John Wiley and Sons, UK. 1989, pp371. Martin, J. (1997), Non-destructive Testing of Metal Ducted Post-tensioned Bridge Beams using Sonic Impact-Echo Techniques, PhD Thesis, The University of Edinburgh. Martin, J. & Forde, M.C. (1995) Influence of Concrete Age and Mix Design on Impulse Hammer Spectrum and Compression Wave Velocity. Construction and Building Materials, Vol. 9, No. 4, 1995, Elsevier Science Ltd. UK Martin, J., Hardy, M.S.A., Usmani, A.S. and Forde, M.C. (1995) Quantifying the defects in posttensioned bridges using impulse ultrasonics. Proc. 6 th Int. Conf.: Structural Faults and Repairs 95. Engineering Technics Press, Edinburgh. Vol 1, Martin, J., Hardy, M.S.A., Usmani, A.S. & Forde, (1998) Accuracy of NDE in bridge assessment, Engineering Structures, Vol 20, No. 11, Martin, J., Giannapolous, A., Hardy, M.S.A. & Forde, M.C. (1999) Ultrasonic Tomography of Impact- Echo NDT of grouted duct P-T RC Beams. Proc. Struct. Faults and Repair 99, Eng. Technics Press. Muldoon, R., Chalker, A., Forde, M.C., Carrera, L. & Kunisue, F. (2003) Identifying voids in plastic p-t ducts in p-t concrete bridge beams - using I-E, SIBIE & tomography, Proc. 10 th Int. Conf. Structural Faults & Repair-20031, Commonwealth Institute, London, 1 st 3 rd July 2003, Engineering Technics Press, CD-Rom, ISBN Papoulis, A. (1977) Signal Analysis, McGraw-Hill, Singapore, 1977, pp431. Petrangeli, M.P. & Sarno, R. (1999) NDT of Prestressing Cables by Impact-Echo method. Proc. Structural Faults + Repair-99, Eng. Technics Press. Poston, R. & Sansalone, M. (1992) Detecting Cracks in the Beams and Columns of a Post-tensioned Parking Garage Structure Using the Impact-Echo Method. Proc. Non-Destructive Evaluation of Civil Structures and Materials, Boulder, Colorado, USA Sansalone, M. & Carino, N.J. (1992) Detection of Voids in Grouted Ducts using the impact-echo Method. ACI Materials Journal. Vol. 92, No. 3, Sansalone, M., Lin, Y. & Carino, N.J. (1990) Finite Element Studies of the Impact-Echo Response of Plates containing Thin Layers and Voids. Journal of Non- Destructive Evaluation. Vol. 9, No. 1, March Sansalone, M. & Street, W. (1995), Use of the Impact- Echo Method and field instrumentation for non- destructive testing of concrete structures. Proc. Inst. Symp. NDT in Civil Engineering (NDT-CE), Berlin, DGZfP, Vol 1, Sansalone, M.J., & Streett, W.B. (1997), Impact- Echo, Nondestructive Evaluation of Concrete and Masonry, Bullpier Press, USA. Stain, R.T. & Dixon, S. (1993) Inspection of cables in post-tensioning bridge - what techniques are available. Proc. 5th International Conference on Structural Faults and Repair - 93, Engineering Technics Press, Edinburgh, UK, Vol. 1, Watanabe, T., Ohtsu, M. & Nakayama Y. (1999) Impact-EchoNDT for grouting performance in P-T ducts. Proc. Struct Faults and Repair-99, Eng. Technics Press. AN

129 Germann Instruments A/S, Copenhagen, Denmark , Impact-Echo Instruments, LLC, Ithica, NY, USA Olson Instruments Inc, Wheat Ridge, CO, USA AN

130 ADVICE NOTE 3.2 SONIC TRANSMISSION AND TOMOGRAPHY FOR MASONRY BRIDGES Contents Chapter 1. Introduction 2. Details of the Technique 3. Tomography 4. Surveying the Structure of Masonry Arches 5. Commissioning and Specification of Sonic Transmission and Tomography 6. Sources of Further Information AN.3.2-1

131 1. INTRODUCTION 1.1 This Advice Note gives guidance on the use of sonic transmission and tomography for the investigation of masonry arch bridges. Further information on defects in masonry arches is provided in Advice Note 2.2 of this series of Advice Notes. AN.3.2-2

132 2. DETAILS OF THE TECHNIQUE Principle of Application 2.1 The direct transmission sonic pulse velocity test (Figure 2.1(a)) involves the passing of a mechanical stress wave through the thickness of the wall or structure under investigation. Transmission of the wave is initiated on one side of the structure and received on the opposite side. The method of initiating and receiving the wave depends on the application for which the method is being used but commonly involves the use of an instrumented hammer and accelerometer (see under Equipment and Procedure below). The resulting wave velocity calculated for a measured path length is an average of the local velocity along the path. The velocity of the sonic compression wave reflects the elastic modulus and density of the materials through which it passes. It cannot pass across air gaps, thus, dry cracks and voids will lead to a longer path with increased transit time and apparently lower velocity. The velocity magnitudes may be plotted in a contour map format, with grid points as X and Y co-ordinates, which allows a simple evaluation of the relative conditions of the material. 2.2 Where two opposite faces are not available, semi- direct or indirect transmission paths may be used (Figure 2.1(b) and (c)) although received energy levels are much lower with a corresponding effect upon interpretation capabilities. Figure 2.1 Transmission Modes for Sonic Wave Tests: a) direct, b) semidirect, c) indirect AN.3.2-3

133 Accuracy 2.3 The accuracy of sonic testing of masonry should be related to the frequency of the input signal and the consequent wavelength of the signal. Work (Martin, Hardy, Usmani, & Forde, 1998) has indicated that the identification of the first detectable target in concrete is at a depth of wavelength/2 (λ/2) and from general geophysics practice it has been assumed that the minimum detectable void size is λ/4. Resolution in masonry is likely to be reduced due to lack of uniformity of the material. Table 2.1 below gives data on typical values of material velocity, hammer mass and hammer input frequency. Of course, different hardnesses of hammer tip in modally tuned hammers may yield different input frequencies on different hardnesses of masonry surface. A modally tuned hammer is one which gives a certain range of frequency input. (The reader is directed to a manufacturer s website for further reading: The contractor/testhouse is asked to give full details of hammers used and anticipated input frequency. Material Velocity (m/s) Hamme r Mass (Kg) Input Frequenc y (Hz) Signal Wavelength, λ (m) Minimum Depth, λ/2 (m ) Minimum Detectable Void Dia λ/4 (m) Material Soil fill 1.4 1, , Poor-average stone 1.4 1, masonry 0.4 1, , Average stone masonry 1.4 1, , , Good stone masonry 1.4 1, , , Good brick masonry 1.4 1, , Table 2.1 Data on Typical Values of material Velocity, Hammer Mass and Hammer Input Frequency Applications 2.4 The method may be used when surveying the structure of masonry arches to identify construction details, voids and poor compaction of fill materials. Advantages 2.5 Since the arrival time of the first wave is of primary concern, no attempt to distinguish complex wave frequencies and reflections is required for the analysis. 2.6 The attenuation of a stress wave is affected by the number and size of flaws and interfaces present in the materials. The level of attenuation of the sonic pulse is less than that of higher frequency signals such as in the ultrasonic pulse test. Attenuation is also related to the number of damped cycles of the propagating wave. The low frequency of the sonic pulse gives fewer cycles over the same distance, resulting in an increase of effective distance over which the measurements can be performed. AN.3.2-4

134 2.7 The technique has a number of potential advantages over other NDT techniques: it is relatively low cost; it is inherently safe compared to techniques such as radiography; it has a greater depth of penetration than ultrasonics or radiography. Limitations 2.8 The survey process is slower than Ground Penetrating Radar, or other methods which can be scanned across the surface, as the receiver has to be temporarily secured to the surface at each test point. 2.9 To determine the pulse velocity, the transit distance is also required at each point Difficulties in calculating the exact time of the wave are a real limitation in sonic tests because of the slightly different characteristics of each recorded sonic pulse, even for the same test location. In some cases, the precise time of initiation and reception of the sonic wave is difficult to define. If transit distances involved are small, minor errors in the determination of the time can thus lead to significant variations of the calculated velocity. The analysis required to determine the transit time where many tests are performed can, therefore, be very time consuming The frequency of the pulse affects the attenuation of the wave and the ability to detect the presence of inhomogeneities. The propagation of the wave can only be affected by flaws which are wider than the wavelength of the pulse. The low frequency associated with the sonic tests results in a longer wavelength than for the ultrasonic case, and the minimum size of inhomogeneity which can affect the transmission of a sonic pulse and hence, be detected is therefore increased Since the measured velocity is the average value along the path, it is not possible to establish the position and extent of any possible inhomogeneity from an individual reading. This may be possible, however, by the use of tomography (see Chapter 3) The case study on which this technique was developed was a near semi-circular arch as found in many parts of Ireland and the UK. However, the technique will be less effective on flat arch bridges with a low rise-span ratio. This is because the sonic signals may well route through the arch rings rather than through the spandrel walls in these latter bridges. Equipment and Procedure 2.14 An instrumented hammer and accelerometer (e.g. Figures 2.2 and 2.3) will typically be used, as transmitter and receiver, together with a multi-channel tape-recorder and dual-channel signal analyser - these could be a two-channel oscilloscope or PC. Measurements are taken at grid points marked on the surface of the structure (e.g. Figure 2.4). It is recommended that a number of readings are taken at each point for averaging (see Section 4.5). AN.3.2-5

135 Figure lb (5.5Kg) large sledge Hammer Figure 2.3-3lb (1.4Kg) short sledge Figure 2.4 Typical Bridge elevation showing grid points for Sonic Testing AN.3.2-6

136 3. TOMOGRAPHY Principle of Application 3.1 The word tomography is derived from the Greek tomos meaning a slice and involves reconstructing a section of an object using measurements taken from outside the object. The tomographic imaging method uses sonic velocity information taken through a section to develop a three- dimensional reconstruction of the velocity distribution in that section. 3.2 Where variations in internal conditions exist, these result in different times of transit being recorded. In particular, the pulse cannot pass across an air gap and must thus, take a longer route leading to an increased transit time. Typical transit paths are given in Figures 3.1 and 3.2. The tomographic software reconstructs the section by combining the information provided by the series of projections obtained at different angles through the element to produce a velocity contour plot. 3.3 Tomography represents an improvement on direct transmission methods because it combines direct and semi-direct methods and because tests are performed in the direct mode also along paths which are not perpendicular to the surface. The section under test is thus crossed by a dense net of transit paths and the tomographic elaboration of the data may account for disrupted wave propagation paths. This gives a detailed map of wave velocity across the structure or sample section so that local values of velocity can be read across the section and the extent and location of flaws can be identified. 3.4 It is usual to assume a linear structural response in tomographic methods. This is because the response is measured with transducers which are normally mounted well away from the location of any impact where non-linearities arise. Any variation from the expected travel time is, therefore, attributed to in-homogeneity or damage in the structure. Accuracy 3.5 The greater the number of readings taken, the more accurate the results (see under Equipment and Procedure below). Applications 3.6 Sonic tomography may be used when surveying the structure of masonry arches. Advantages 3.7 The technique significantly enhances interpretation capabilities compared with that of a series of individual readings. AN.3.2-7

137 Figure 3.1 Tomographic Paths Figure 3.2 Ray Path Coverage for Different Transducer Arrangements Limitations 3.8 The procedure is relatively slow and requires specialist software for interpretation. Equipment and Procedure 3.9 Equipment will be as for individual measurements, supplemented by specialist software for tomographic reconstruction. It may be convenient to use multiple receivers fixed to the masonry surface to speed operations and reduce the number of hammer blows required at any one position (Figure 3.3). Measurements will be taken between adjacent or opposite surfaces as shown in Figures 3.1 and 3.2. AN.3.2-8

138 3.10 It is important that transit paths are uniformly distributed across the section investigated and that the maximum possible area is crossed by the sonic paths. In this respect, Figure 3.4(b) represents an improved configuration over the section represented in Figure 3.4(a). Both cases have been obtained by locating 12 reading stations around the section, but the modified spacing of case (b) not only allows for better coverage of the area but the number of transit paths also increases from 92 to 104 (or 208 if back readings are recorded, to account for anisotropy of the material) A number of inversion algorithms are available for tomographic reconstruction. Figure 3.3 Receiving transducers fixed to masonry arch for sonic tomography Shear wave geophones showing attachment detail to structure by means of plates and rawl bolts Figure 3.4 Alternative Transit Patterns AN.3.2-9

139 4. SURVEYING THE STRUCTURE OF MASONRY ARCHES Method of Application 4.1 The methods can be used to comparatively assess the internal condition of masonry arch bridges, including abutments. Further information on defects in masonry arches is provided in Advice Note 2.2 of this series of Advice Notes. 4.2 In most cases, either the direct or semi-direct transmission method will be used (see Chapter 2 under Principle of Application). If the direct method is used, the receiver will normally be located directly opposite the point of impact for individual readings. Tomography procedures are outlined in Chapter The frequency of the vibrations in the structure resulting from the propagating wave is low, generally around 1kHz. The resulting wave velocity in a masonry arch is an average of the local velocity along the path of brick or stone, mortar joints, soil fill and air voids or other defects. Since sonic waves cannot pass across an air gap, a propagating wave must find a path around an air-filled void, resulting in both attenuation and an increase in transit time of the signal. Accuracy 4.4 See Chapter The accuracy of the velocity reconstruction can be maximised by a carefully planned choice of position and number of the reading stations and by simple data smoothing prior to analysis. A typical grid of test points would be of the order of 0.5m to 1.0m spacing - both vertically and horizontally. Applications 4.6 To determine the presence of flaws, inhomogeneities and poor density areas when access is available to more than one face of the structure. Repair and retrofit procedures such as grout inspection can also be assessed by tomography, by examination of velocity contour plots. 4.7 Flaws could be in the form of a crack, void or delamination at the interface between brick or stone and mortar, or an interface between wythes. Applications include assessment for detection and location of voids and unconformities in the fill and masonry, via direct transmission tests through the spandrel walls. It may be possible to estimate the depth of surface cracks from indirect measurements. Advantages 4.8 The method non-destructively provides data relating to the interior of masonry arch construction. Good depth penetration and comprehensive coverage are possible. The position of features can be estimated from tomography, which is not possible with individual readings. 4.9 Sonic waves are elastic deformation waves and their propagation characteristics (velocity and attenuation) depend on the elastic properties of the materials they cross and the anisotropic behaviour of the masonry. Anisotropic effects can be studied by measuring sonic wave velocities along opposite directions of the structure. AN

140 Limitations 4.10 In field measurements, high noise levels can be expected from both the physical measurement and the electronic instrumentation used for data capture and processing. Vibrations produced by traffic may also generate noise on the received signal. Therefore, averaging of a number of repeated readings at the same position is generally a requirement before the transit time is calculated In most masonry walls or structures, the thickness is fairly consistent, varying only by one or two per cent. The exact thickness is, however, difficult to establish at each test point. In this case, average values from all accessible locations are used Data acquired usually exhibit a good deal of velocity scatter, resulting from variations in the strength and nature of the hammer blow generating the input signal, the interpretation of acquired waveforms by the operator and coupling of the receiving transducer to the masonry surface. The measured velocity will also be influenced by moisture content, which may not be uniform. Data scatter has the effect of increasing the residual error of velocity reconstruction and may lead to identification of false anomalies Typical defect sizes identifiable depend on the material and hammer input frequency see Table It should be noted that the first received signal peak may not have followed a direct path from transmitter to receiver but may have followed a more tortuous route through materials with higher velocity. Structure geometry may, thus, be an important factor, especially where tomography is used. This may be particularly important when testing flat arch structures Some surface damage may also be caused, depending upon hammer size and nature of the masonry, by the need for repeated readings at the same location. Care should be taken to minimise this as far as is possible. The tomographic procedure may be very time consuming unless multiple receiving transducers are used. Equipment and Procedures 4.16 As outlined in Chapter 2 under Equipment and Procedure, the vibrations in the structure resulting from the propagating wave induced by an instrumented hammer are measured by an accelerometer mounted on the masonry surface with an appropriate acoustic coupling medium such as grease. Fixing by means of bolts in holes drilled into mortar beds may be appropriate where tomography is being used (see Figure 3.3). (It is important that good transducer coupling is achieved at the receiving station). A two dimensional grid is marked out on the surface and a test is performed at each node point. The signals of both the hammer and the accelerometer are recorded by the acquisition system for each point. The time interval for the sonic wave to travel between the hammer and the accelerometer is defined as the transit time. The data collected from the test provides the transit time of the stress waves through the path connecting transmitter (hammer) and receiver (accelerometer) across the structure A transmission distance of up to about 10m may be achieved with a 1.8 kg hammer and 20m with a 5.5kg hammer. See Figure 4.1: AN

141 Figure 4.1 Typical Sonic Transmission arrangement 4.18 There is a marked reduction in measured velocity as the angle between transmitter and receiver is increased. This effect is due to the anisotropic properties of the masonry materials, which prevent the wave from propagating with a spherical migrating wavefront. This anisotropic material behaviour must be accounted for in tomographic reconstruction by calculating the ratio between V d and V s V d /V s where V d = direct path velocity V = semi-direct path velocity (transmitting and receiving transducers on adjacent surfaces or not directly opposite) which decreases with increasing incidence angle. The minimum averaged value of semidirect velocity will be used in the ratio Incidence angles approaching 90 o should be avoided, as should very short transit paths. It is advisable to increase the number of transit paths to increase data redundancy as discussed in Chapter 3 under Equipment and Procedure above. Spurious readings (those showing isolated high deviations from the average velocity which are outside/greater than the variation known to be produced by any defects, etc), should be eliminated prior to inputting to the tomographic software It is recommended that identified anomalies should be further investigated by drilling or excavation. Interpretation 4.21 To analyse the results, the pulse velocity of the wave can be easily calculated at each location by the equation v = d/t where v = pulse velocity d = length of path t = transit time Typical values of direct transmission pulse velocity in a range of materials are reported in Table 4.1. For further interpretive information, see under Interpretation Notes below. AN

142 Material Velocity of compression wave (m/s) Air 330 Water 1500 Concrete Brick 3400* Sandstone , 2100* Steel 5900 Fill material (clay soil, etc) * Values measured in the field Table Compressional Wave Velocities through a Selection of Construction Materials 4.23 The travel times of individual readings should be carefully studied and an accuracy assessed. (This may typically be in the range of ± 5 to ± 10% depending on path length and equipment used.) 4.24 Interpretation of tomography results is based on visual inspection of the velocity plots for sections through the structure. High velocity areas located along the edges can represent stone masonry walls, whereas medium and low velocity areas located internally across the plot can represent fill materials or voidage or cracks (See also under Chapter 2, Accuracy) As indicated under Equipment and Procedures above, it is recommended that interpretation should be confirmed by drilling or excavation at representative locations. Example 4.26 North Middleton Bridge, in Lothian, Scotland, was selected on the basis of its accessibility and overall characteristics and dimensions Reasons for testing were as follows: sonic tests had previously been performed on the east abutment/wing walls of the bridge and the new tests provided a chance to monitor the structure after maintenance work and also permitted comparison of the outcome of the structural evaluation obtained through different NDT techniques and through sonic tests applied in different modes; a study of the performance of the techniques would consider the effectiveness of the testing methods as a function of the site conditions in order to identify requirements for useful applications North Middleton Bridge is a twin span arch with each arch spanning 3.7m and the width of the bridge being just over 8m - see Figure 4.2a. Non-destructive sonic surveys were undertaken on the east abutment and wing walls, which are shown on the left hand side of the Figure. AN

143 a) Bridge elevation (White line marks level of radar and sonic tomography) b) Main dimensions and wing wall where sonic tests were performed Figure North Middleton Bridge, downstream view AN

144 4.29 Both individual transmission and tomographic measurements were made, with tomography at one particular level (also used for radar tomography) as shown in Figure 4.2a Individual tests were carried out on the wing walls and abutment of the bridge as Figure 4.2b using a modally tuned impact hammer, fitted with a force transducer, to excite the structure. An accelerometer was fixed to the wall to monitor the structure response waveforms. The hammer and accelerometer were connected to a multiple channel tape recorder, so that two simultaneous channels of data were recorded digitally on tape for post processing with a dual channel 25kHz signal analyser. An alternative system would involve using a PC with A/D cards fitted. The hammer used was a PCB 12lb large sledge with a modally tuned black plastic tip giving an input frequency of 500Hz ( It is essential to use such a large low frequency hammer in order to achieve depth penetration. A metal tipped hammer of such a mass would cause substantial damage to the masonry surface and also generate a higher frequency which would be attenuated quickly Readings were collected at a number of horizontal levels on the structure under investigation by marking a regular grid on the walls. At each level measurement points were located on the abutment face, downstream wing wall and upstream wing wall at 1 metre spacings. All points on the grid were alternatively used as transmitting and receiving stations, to allow collection of back readings The waveforms obtained were analysed in the time domain. The time delay between the excitation given with the hammer and recorded on the first channel, and the reception of the hammer blow on the second channel, together with the distance between source and receiver, permitted calculation of the velocity of the sonic signal through the material. This value is dependent upon the quality of the material (in this case stone masonry and soil fill) and can be used to rank the integrity of the structure More detailed discussion is contained in Forde & Batchelor, 1984 and Forde, Komeyli-Birjandi and Batchelor, Tomographic testing used the same equipment at level 2m of the abutment as shown in Figure 4.2a. Fifteen stations were used for a total of 142 transits. Figure 4.3 shows a horizontal cross-section of the abutment with the locations of the transmitters and receivers (Transit paths were assumed to be straight at the beginning of the tomographic reconstruction). AN

145 Figure 4.3 North Middleton Bridge - Plan of abutment and wing walls (2m level): initial assumed density of sonic ray paths 4.35 From the first data analysis (Figure 4.4) it is clear that high velocity areas are located along the walls corresponding with the stone masonry, while values of velocity between medium and low are situated towards the centre of the plot, indicating backfill materials. The bridge was identified as a cellular construction, hence, the low transmission velocities in the centre of cross- sections Looking in more detail, along the upstream and especially downstream walls, very high velocity areas relate to a continuous homogeneous dense material that subsequent endoscopic investigations revealed to be cement grouted zones. By applying calibrations, such as wall thickness calibration from drilling, to the model and reiterating (Figure 4.5), masonry thickness on the downstream side was calculated and found to be comparable with measurements determined from other tests. Thus, calibration was an important part of the survey. AN

146 Figure Sonic unconstrained data Figure Sonic constrained data Note: Unconstrained data - assumes no pre-knowledge of construction. Constrained data - permits input of known facts, e.g. estimate of wall thickness. Documentation 4.37 Details of the relevant theory and descriptions of the development and use of the techniques are given in the above references, and in the sources identified in Chapter 6. Interpretative Notes 4.38 Typical values of sonic velocity through individual materials are indicated in Table 2.1, whilst reported values averaged thorough spandrel walls and soil fill are generally of the order of 1000 to 1500m/s. One difficult point to establish is the pulse velocity through an unflawed region of masonry, with stone masonry in the range 1,500-2,500m/s and brick masonry 2,500-3,500m/s. In addition, variations in masonry material components may have some effect on wave propagation. Direct comparison between two different material types is, thus, difficult to make. However, evaluation of the relative uniformity of a given material can be effectively performed Aspects to be considered in interpretation are as follows: Depth of Propagation into the Medium 4.40 The attenuation of sonic waves is governed by absorption due to frictional losses and spherical spreading The ability to detect an internal discontinuity is dependent on the amplitude of the reflection, see under Reflection and Transmission of Energy below. This means that if a very high proportion of energy is reflected between different media then little energy will penetrate deeper than the reflector. In the extreme case for compressional waves. virtually no energy will pass through air, i.e. complete reflection occurs, for example, at an air gap or dry void. Reflections from embedded metals will be less marked. AN

147 Signal Velocity 4.42 Wave impulses are propagated with a velocity, as explained above, which is dependent on the material of propagation. In the case of sonic waves there are two possible modes of vibration through a material; compressional waves and shear waves. The waves which travel through a material as the result of a hammer blow are mainly in the form of compressional waves. The factors which control velocity are the elastic constants and density. Reflection and Transmission of Energy 4.43 An impulse of sonic energy is partially reflected and partially transmitted at a boundary between two materials as a result of differences in the acoustic impedance of the materials. The acoustic impedance is found from the product of density and velocity. The greater the contrast in the impedance the stronger the reflections and the less the amount transmitted through the boundary This is demonstrated by the expressions for reflection and transmission coefficient 4.45 Reflection Coefficient and Transmission Coefficient where Z1 = impedance of first material Z2 = impedance of second material Coupling to the Surface 4.46 Coupling refers to the ability to transfer energy into or out of the material under investigation. Mechanical impact techniques, by definition, require direct contact with the surface of the material. In the case of the receiving accelerometer, a coupling agent such as petroleum jelly or grease is normally used for this purpose (see under Equipment and Procedures above). Variations in coupling efficiency will lead to scatter of results and possible misinterpretation. AN

148 5. Commissioning and Specification of Sonic Transmission and Tomography Introduction 5.1 The following requirements are specific to the use of sonic transmission and tomography for the investigation of masonry arches and should be used in conjunction with those given in Advice Note 1 General Guidance of this series of Advice Notes. Information Supplied to Tenderers 5.2 The tender documents should require the following additional information to be provided: the number of sonic transmission transits required through each element and whether tomography is to be used; all available construction details of the structure to be tested. Information Required of Tenderers 5.3 The tenderers should be asked to state: the type of sonic survey proposed and proposed transit locations; details of the hammer to be used. Report 5.4 The Specification should require the following information to be included: details of the sonic survey carried out and the transit locations used. AN

149 6. SOURCES OF FURTHER INFORMATION Colla, C. (1997) Non-destructive testing of masonry arch bridges, PhD thesis, University of Edinburgh, p272. Davidson, N. & Forde, M.C. (1998) Field measurements of sonic velocity, University of Edinburgh internal report. Forde, M.C. & Batchelor, A.J. (1984) Low Frequency NDT testing of historic structures, Proc. 3 rd European Conf on NDT, Florence, Oct 1984, Vol 3, Forde, M.C., Komeyli-Birjandi, F. & Batchelor, A.J. (1985) Fault detection in stone masonry bridges by non- destructive testing, Proc Int. Conf. Structural Faults & Repair-85, April 1985, London, Engineering Technics Press, Jackson, M.J. & Tweeton, D.R. (1994) MIGRATOM, Geophysical Tomography Using Wavefront Migration and Fuzzy Constraints, R.I. 9497, Bureau of Mines, United States Department of Interior, 35 pp. Jalinoos, F., Olson, L.D. & Sack, D.A. (1995) Use of Combined Acoustic Impact Echo and Crossmedium Tomography Methods for Defect Characterisation in Concrete Civil Structures, Proc. 6 th Int. Conf. On Structural Faults and Repair London, Engineering Technics Press, Vol. 2, Kak, A.C. & Slaney, M. (1988) Principles of Computerised Tomographic Imaging, The Institute of Electrical and Electronic Engineers Press, New York. Kingsley, G.R., Noland, J.L. & Atkinson, R.H. (1987) Nondestructive Evaluation of Masonry Structures using Sonic, and Ultrasonic Pulse Velocity Technique, Proc. 4 th North American Masonry Conference, Los Angeles, CA, August Komeyli-Birjandi, F., Forde, M.C. & Batchelor, A.J. (1987) Sonic analysis of masonry bridges, Proc. 3 rd Int. Conf. Structural Faults & Repair-87, University of London, July 1987, Vol 2, Engineering Technics Press, Martin, J., Hardy, M.S.A., Usmani, A.S. & Forde, M.C. (1998) Accuracy of NDE in bridge assessment, accepted for publication in Journal: Engineering Structures, Vol. 20, No. 11, Schuller, M., Berra, M., Atkinson, R. & Binda, L. (1995) Acoustic Tomography for Evaluation of Unreinforced Masonry, Proc, 6 th Int. Conf. Structural Faults and Repair 95, Vol. 3, Sibbald, A. & Forde, M.C. (1987) Investigation of brick masonry in man-entry sewers, Proc. 3 rd Int. Conf. Structural Faults & Repair-87, University of London, July 1987, Vol 2, Engineering Techniques Press, AN

150 ADVICE NOTE 3.3 ULTRASONIC TRANSMISSION AND TOMOGRAPHY FOR POST- TENSIONED CONCRETE BRIDGES Contents Chapter 1. Introduction 2. Details of the Ultrasonic Technique 3. Tomography 4. Assessing the Grouting of Post-Tensioned Concrete 5. Commissioning and Specification of Ultrasonic Transmission and Tomography 6. Sources of Further Information AN.3.3-1

151 1. INTRODUCTION 1.1 This Advice Note gives guidance on the use of ultrasonic transmission and tomography for detecting voids in grout in post-tensioned tendon ducts in prestressed concrete bridge elements. Further information on defects in the grouting of ducts in post- tensioned concrete is given in Advice Note 2.1 of this series of Advice Notes. AN.3.3-2

152 2. DETAILS OF THE ULTRASONIC TECHNIQUE Principle of Application 2.1 The time for a pulsed compression wave with a frequency typically in the range 25kHz to 100kHz to pass between transmitting and receiving transducers located on opposite faces of a concrete element is measured. This is known as the Direct Transmission method and is the most reliable approach as illustrated in Figure 2.1. Alternatively, the transducers may be located on adjacent faces as also illustrated in Figure 2.1 (Semi-Direct Transmission). The pulse velocity may be calculated from knowledge of the path length between the transducers, which must also be measured, as described in Chapter 4 under Interpretation. 2.2 The pulse velocity is dependent on the density and elastic modulus of the concrete, and will be influenced by moisture conditions. The waves cannot pass through air, and will thus, pass round airfilled cracks, voids and honeycombed concrete. This will lead to a greater path length as illustrated in Figure 2.1, and hence, increased transit time and lower apparent velocity. 2.3 Direct transmission measurements are commonly taken on a grid of points at appropriate spacings located around the area of interest with results plotted as contours of either transit time or velocity to facilitate interpretation. Figure 2.1 Ultrasonic Transmission Modes and Effects of Air Void AN.3.3-3

153 Accuracy 2.4 Estimates of pulse velocity will be influenced by the accuracy with which the path length can be measured. Accuracies of pulse velocity should be possible to within ± 2%. Void detection will depend upon grid spacing, and voids smaller than this are unlikely to be detected. A path length increase greater than 2% will also be necessary. 2.5 More detailed guidance is given in Chapter 4 under Accuracy. Applications 2.6 The method may be used when surveying concrete elements to identify variations in material properties and the presence of internal defects. Advantages 2.7 Measurement of the velocity of ultrasonic pulses in concrete is a well-established technique for site application using commercially available equipment. The method is relatively low cost, straightforward to use, and there are no specific safety hazards. Limitations 2.8 The signal will be heavily attenuated by the presence of aggregate particles, which will limit path lengths according to frequency with consequent effects on resolution of internal features (see Chapter 4 under Accuracy). The velocity of the signal in reinforcing steel is greater than in concrete, and this should be avoided wherever possible. It will normally be necessary to use a suitable couplant between the transducers and the concrete surface which will slow the testing process and may also leave surface staining in some cases. Further limitations are identified under Accuracy above and in Chapter 4 under Limitations. Equipment and Procedure 2.9 Portable battery-powered digital equipment is widely available to generate and receive suitable signals and to indicate the time delay caused by passing through the concrete. A range of hand-held transducers is also available. Further information is provided in Chapter 4 under Equipment and Procedures. AN.3.3-4

154 3. TOMOGRAPHY Principle of Application 3.1 The word tomography is derived from the Greek tomos meaning a slice and involves reconstructing a section of an object using measurements taken from outside the object. The tomographic imaging method uses ultrasonic pulse velocity information taken through a section to develop a two or three-dimensional reconstruction of the velocity distribution in that section. 3.2 Where variations in internal conditions exist, these result in different times of transit being recorded. In particular, the pulse cannot pass across an air gap and must thus take a longer route leading to an increased transit time. Typical transit paths are given in Figures 3.1 and 3.2. The tomographic software reconstructs the section by combining the information provided by the series of projections obtained at different angles through the element to produce a velocity contour plot. 3.3 Tomography represents an improvement on direct transmission methods because it combines direct and semidirect methods and because tests are performed in the direct mode also along paths which are not perpendicular to the surface. The section under test is thus crossed by a dense net of transit paths and the tomographic elaboration of the data may account for disrupted wave propagation paths. This gives a detailed map of wave velocity across the structure or sample section so that local values of velocity can be read across the section and the extent and location of flaws can be identified. 3.4 It is usual to assume a linear structural response in tomographic methods. This is because the response is measured with transducers which are normally mounted well away from the location of any impact where non- linearities arise. Any variation from the expected travel time is, therefore, attributed to inhomogeneity or damage in the structure. Accuracy 3.5 The greater the number of readings taken, the more accurate the results. Applications 3.6 Ultrasonic tomography may be used when assessing the grouting of post-tensioned construction. Advantages 3.7 The technique potentially enhances interpretation capabilities compared with that of a series of individual readings. Limitations 3.8 The procedure is slow, requires specialist software for interpretation, and is not yet fully developed commercially. Equipment and Procedure 3.9 Equipment will be the same as for individual measurements, supplemented by specialist software for tomographic reconstruction. Measurements will be taken between adjacent or opposite surfaces as shown in Figures 3.1 and 3.2. AN.3.3-5

155 3.10 It is important that transit paths are uniformly distributed across the section investigated and that the maximum possible area is crossed by the paths. In this respect, Figure 3.3(b) represents an improved configuration over the section represented in Figure 3.3(a). Both cases have been obtained by locating 12 reading stations around the section, but the modified spacing of case (b) not only allows for better coverage of the area but also the number of transit paths increases from 92 to 104 (or 208 if back readings are recorded, to account for possible anisotropy of the material) A number of inversion algorithms are available for tomographic reconstruction. Figure 3.1 Tomographic Paths Figure 3.2 Ray path Coverage for Different Transducer Arrangements AN.3.3-6

156 Figure 3.3 Alternative Transit Patterns AN.3.3-7

157 4. ASSESSING THE GROUTING OF POST- TENSIONED CONCRETE Method of Application 4.1 The method seeks to identify voids in the tendon ducts of post-tensioned concrete using measurements of ultrasonic pulse velocity. It can be used to aid interpretation of Impact Echo results, and to resolve ambiguous results from an Impact Echo survey. Further information on defects in the grouting of ducts in post- tensioned concrete is given in Advice Note 2.1 of this series of Advice Notes. 4.2 The velocity of an ultrasonic compression wave pulse in concrete primarily reflects the elastic modulus and density of the material. Low velocities indicate poor quality material whilst air-filled cracks and voids will lead to a longer path and hence greater measured transit times and lower apparent velocities. 4.3 Direct transmission measurements may be taken between opposite faces of the member, such that the tendon duct lies directly on the path between transmitting and receiving tranducers, at positions along the length of the beam. Alternatively, a tomographic approach may be used as described below. Accuracy 4.4 This will be influenced by the spacing between test positions as well as the sensitivity of the equipment and analysis techniques used. A tomographic grid spacing of 75mm vertically and 80mm horizontally has been used experimentally but this was not a sufficiently fine grid to allow detailed investigation of 40mm diameter ducts. (see below). However the tomographic contour plot did appear to indicate the location of ducts that size. Another experiment used 100mm grid spacing and this indicated voids in ducts believed to be 100mm in diameter (see below). Smaller spacings may be necessary in practice. 4.5 Resolution is unlikely to be better than half the wavelength (λ/2 is typically about 40mm in good quality concrete for a commonly used 54kHz pulse frequency, or 80mm for a 25kHz frequency or 25mm for a 62kHz transducer) and may be worse. These values reflect the minimum depth and size of detectable feature whilst the least internal dimension of the element should exceed λ. 4.6 Practitioners need to be aware that new multi-transducer array ultrasonic systems are starting to become available some of these systems use shear wave transducers, giving greater depth penetration. Little information has been published on their applications. Applications 4.7 Where the velocity contour plot for a posttensioned element shows a clear reduction in velocity at the location of duct, it is likely that the duct is voided at that point. A low velocity area of the element can also indicate poor quality concrete. A typical plot is shown in Figure 4.1, where the grey-scale indicates zones of different pulse velocity. Advantages 4.8 The method can provide information relating to the interior of a post-tensioned beam. AN.3.3-8

158 Limitations 4.9 If access is not possible all-round the perimeter of the element, for example, the sides and soffit of a beam may be accessible, but possibly not the top, there will be limitations on the coverage of the beam near the top where using tomography. Edge effects, where there is a reduction in the number of transit paths, may also limit the coverage close to the surface. Tomography is particularly sensitive to other aspects of member geometry such as non-parallel surfaces of flanges, duct position and measurement accuracy. Receiver positions should be spread as widely as possible to seek to minimise these effects. Testing speed may be limited by the need, with some equipment, to provide coupling between the transducers and concrete surface for which petroleum jelly or similar is commonly used Using a practical grid size may result in difficulties in resolving two ducts lying adjacent to each other since these may appear as a single feature. The presence of dense reinforcing steel around tendon anchorage zones near beam ends may also cause difficulties The transit times for both direct transmission and tomography can be measured using an ultrasonic digital tester, such as widely used in non-destructive testing of concrete. A test grid is set up at the section to be investigated using similar distances between test points on all test surfaces. Each transducer location and corresponding time of transit is noted and subsequently input into a model to be analysed by appropriate tomographic modelling software where this is being used. The internal wave velocities calculated are then presented as the contour plot for the section. It is recommended that accurate duct positions should be established using an appropriate technique (such as GPR) prior to establishing ultrasonic test points. Direct transmission readings will be considerably quicker than tomography. A series of initial measurements at positions along the length of the duct may focus attention on localised areas for more detailed tomographic study which may be able to differentiate between voiding in the duct and honeycombed concrete. It is recommended that potential voids or other anomalies are confirmed by drilling and endoscopic examination. Interpretation 4.12 To analyse the results, the pulse velocity can easily be calculated at each location by the equation: v = d/t where v = pulse velocity d = path length t = transit time 4.13 Typical values for concrete range from about 3000 to 4500m/s depending on quality and moisture condition For assessing the grouting of ducts in post- tensioned concrete the following should be considered: dense reinforcement at the ends of the beams (near the anchor points) may affect the accuracy of the readings; or outer ducts may mask the readings from inner ducts; mm spacing of links should be less problematical; localised drilling into ducts (with care not to damage the tendons) should be used to confirm the presence of voiding and to confirm absence of voiding at selected critical locations. AN.3.3-9

159 4.15 Where equipment permits it is recommended that individual signal waveforms (A-Scans) obtained from direct transmission readings, at adjacent transit paths should be stacked along the length of the beam under investigation to form a B-Scan plot. This will enable anomalous zones to be more readily identified Reduced velocity at the location of a duct implies voiding. In one trial Impact Echo testing had indicated a reflection in the vicinity of a duct. The tomographic survey showed normal velocity through the duct, but reduced velocity below the duct, suggesting that there might be voiding in the concrete immediately below the duct. As stated above, reduced velocity can mean poor concrete locally within the beam. Examples Case Study at TRL 4.17 A tomographic survey was carried out at a section of a beam 750mm deep by 400mm wide with a cross-section as shown in Figure 4.1. Grid locations were spaced 75mm vertically and 80mm horizontally. The 40mm diameter ducts could not be investigated in detail, but possible voiding in ducts was discernible (see Figure 4.2). Case Study at Stanger Science and Environment 4.18 Tomographic surveys were carried out at three sections on a beam 750mm deep by 400mm wide with grid locations at 100mm spacing - using all four faces. The ducts are believed to be 100mm in diameter. Voiding in one duct was clearly shown by Tomography and confirmed by Impact Echo. Apparent areas of low velocity in the corners are due to the reduced number of transit paths producing unreliable results. (See Figures 4.3, 4.4 and 4.5). These results may be compared to the as constructed beam in Figure 4.3. Note: Edge effects or errors occur where there is a corner or low density of transmission/reception rays in a model. Edge effects or errors are highlighted in the cross-sections. AN

160 Figure 4.1 Post-tensioned TRL Beam AN

161 km/sec NOTE: Contours show areas of high and low velocity as per the key. Low velocity indicates voiding/poor concrete. High velocity represents good unvoided concrete. Figure TRL Tomographic survey - Position 4.4m from track end AN

162 km/sec NOTE: Contours show areas of high and low velocity as per the key. Low velocity indicates voiding/poor concrete. High velocity represents good unvoided concrete. Figure Stanger Science & Environment beam Tomographic Survey Position 0.4m from front end AN

163 km/sec NOTE: Contours show areas of high and low velocity as per the key. Low velocity indicates voiding/poor concrete. High velocity represents good unvoided concrete. Figure Stanger Science & Environment beam Tomographic Survey. Position 0.8m from front end AN

164 km/sec NOTE: Contours show areas of high and low velocity as per the key. Low velocity indicates voiding/poor concrete. High velocity represents good unvoided concrete. Figure Stanger Science & Environment beam Tomographic Survey. Position 0.9m from front end Documentation 4.19 Background to the development of the technique is provided in Chapter 6. AN

165 5. COMMISSIONING AND SPECIFICATION OF ULTRASONIC TRANSMISSION AND TOMOGRAPHY Introduction 5.1 The following requirements are specific to the use of ultrasonic transmission and tomography for assessing the grouting of post-tensioned tendons and should be used in conjunction with those given in Advice Note 1 General Guidance of this series of Advice Notes. Information Supplied to Tenderers 5.2 The tender documents submitted to testing organisations should require the following additional information to be included: as-built drawings providing reinforcing details, duct types and sizes, profiles, locations and depth of cover; which ducts and which lengths of ducts are to be tested in detail; whether other ducts or intermediate lengths of the above ducts are to be tested at sample locations. Interval to be stated. 5.3 The following additional deliverables from the testing organisations should be stated: for each test location whether the techniques used have been able to detect a void in the duct; the minimum size of void that it would have been possible to detect; whether any ducts were too close to the surface for voids to be detected. Information Required of Tenderers 5.4 The Tenderers should be asked to state the following additional information: proposed transit locations. Report 5.5 The Specification should require the following additions to the report. a discussion of the number of transits measured at each cross-section and a demonstration of whether these were adequate to locate structural features or defects at all locations required; raw ultrasonic traces and, if applicable, filtered ultrasonic traces to provide a clear view of the quality of the data print out; a sensitivity analysis indicating the minimum depth which can be resolved. AN

166 6. SOURCES OF FURTHER INFORMATION BS1881: Part 203. Recommendations for the measurement of ultrasonic pulses in concrete. British Standards Institution, London. Cavell, D.G. (1997) Assessment of deteriorating post- tensioned concrete bridges, PhD thesis, University of Sheffield. Colla, C. (1997) Non-destructive testing of masonry arch bridges, PhD thesis, University of Edinburgh, p272. Jackson, M.J. & Tweeton, D.R. (1994) MIGRATOM, Geophysical Tomography Using Wavefront Migration and Fuzzy Constraints, R.I. 9497, Bureau of Mines, United States Department of Interior, 35 pp. Martin, J., Broughton, K.J., Giannopoulos, Hardy, M.S.A. & Forde, M.C. (2001) Ultrasonic Tomographic Impact-Echo NDT of Grouted Duct P-T R.C. Beams, NDT&E International, Elsevier Science, 2001, 34, Martin, J. & Forde, M.C. (1995) Influence of concrete age and mix design on impulse hammer spectrum and compression wave velocity. Construction and Building Materials 1995, 9, No. 4, Martin J., Giannapolous A., Hardy M.S.A. & Forde M.C. (1999) Ultrasonic Tomography of Impact-Echo NDT of grouted duct P-T RC Beams. Proc. Struct. Faults and Repair 99, Eng. Technics Press CD-Rom. Martin, J., Hardy, M.S.A., Usmani, A.S. & Forde, M.C. (1995) Quantifying the defects in posttensioned bridges using impulse ultrasonics. Proc. 6th Int. Conf.: Structural Faults and Repairs 95. Engineering Technics Press, Edinburgh. Vol 1, Martin, J., Hardy, M.S.A., Usmani, A.S. & Forde, (1998) Accuracy of NDE in bridge assessment, Engineering Structures, Vol 20, No. 11, Stain, R.T. & Dixon, S. (1993) Inspection of cables in post-tensioning bridge what techniques are available. Proc. 5th International Conference on Structural Faults and Repair 93, Engineering Technics Press, Edinburgh, UK Vol 1, AN

167 ADVICE NOTE 3.4 ELECTRICAL CONDUCTIVITY Contents Chapter 1. Introduction 2. Details of the Technique 3. Surveying the Structure of Masonry Arches 4. Commissioning and Specification of Electrical Conductivity 5. Sources of Further Information AN.3.4-1

168 1. INTRODUCTION 1.1 This Advice Note gives guidance on the use of Electrical Conductivity Measurements for the investigation of Masonry Arch Bridges. 1.2 The purpose of the technique is to identify changes in relative conductivity, rapidly and at low cost. These changes can then form the basis for further investigation. 1.3 This test method covers the determination of moisture behind the surface, of voids and of mineral concentration in masonry arch bridges using electrical conductivity measurements. The electrical conductivity method enables low cost scans to be made of the sub- surface conductivity of the bridge fabric with a nominal penetration of normally 1.5m or, exceptionally, up to 6m where large surface areas are available (and a large instrument can be used). This NDT method requires no contact with the surface of the structure and is non- invasive. AN.3.4-2

169 2. DETAILS OF THE TECHNIQUE Principle of Application 2.1 The reciprocal of the electrical resistivity is defined as the electrical conductivity, a measure of the ease with which an electrical current can be made to flow through a substance. In the MKS system the unit of conductivity is the mho per metre or Siemen per metre and a resistivity of one ohmmetre (1 Ωm) exhibits a conductivity of one mho per metre (1 ω/m) or one Siemen per metre (1 S/m). For convenience, conductivity values are usually defined in millisiemens per metre (ms/m). 2.2 The application of this electromagnetic technique for measuring conductivity involves the use of a transmitter coil energised with an alternating current and a receiver coil located a short distance away. The instrument produces a constantly changing magnetic field and this field induces eddy currents within the tested object according to the conductivity of the component materials. These eddy currents produce a constantly changing magnetic field, which, in turn, induces a current in the receiver coil. 2.3 The primary field also induces a current in the receiver coil. As the eddy current takes time to generate, however, a phase lag occurs between the two induced currents in the receiver coil. The amount of phase lag is dependent on the conductivity of the material and the frequency used. Furthermore, the primary electromagnetic field creates induced magnetisation in magnetic materials which in turn produce a secondary field in the receiver coil that is different from the eddy current fields discussed above. The received signal, therefore, has two components: the out of phase, or quadrature component, which is mostly related to the material/s conductivity, and the in-phase component, which is an indicator of the relative magnetic susceptibility of the tested object. 2.4 The secondary field is a function of the inter-coil spacing, the operating frequency and the conductivity of the materials. It can reveal the presence of a conductor and provides information on its geometry and electrical properties. The electrical conductivity of the medium is a function of several factors. These include porosity, shape of porosity, degree of saturation, temperature and presence of clays with moderate to high cation exchange capacity in any fill material. Clearly, metalliferous objects buried within the near- surface will contribute to the bulk electrical conductivity of the medium. 2.5 The induction of current flow results from the magnetic components of the electromagnetic (EM) field, and consequently there is no need for physical contact with the surface of the structure investigated. A widely used commercially available conductivity meter is the Geonics EM 38 Figures 2.1, 2.2 & 2.3. By measuring the quadrature component, the Geonics EM38 instrument can directly measure the apparent conductivity of the ground down to a maximum depth of about 1.5 metres with an accuracy of approximately 5% 1. 1 Geonics Ltd. (1992) Geonics EM38 Ground Conductivity Meter Operating Manual Geonics Ltd: Ontario, Canada AN.3.4-3

170 Figure Showing the Horizontal Position (parallel to investigated surface, surface assumed to be a vertical wall) and Vertical Position orientation (normal to investigated surface, surface assumed to be a vertical wall) Figure When the instrument is held perpendicular to the measured surface (termed vertical orientation) the measured signal is representative of the condition of the structure between 0.1m and 1.5m from the surface. In this orientation peak sensitivity is at a depth of c.400mm. Note that although the instrument is almost horizontal, it is still in vertical orientation relative to the surface Figure When the instrument is held parallel to the measured surface (termed horizontal orientation), the measured signal is representative of the condition of the structure between the surface and 1m depth. In this orientation peak sensitivity is at the surface. Note that while the instrument is almost vertical, it is still in horizontal orientation relative to the surface Applications AN.3.4-4

171 2.6 The technique can be used for indications of water ingress and moisture movement into structures (Chapter 3 Paragraph 4). The output from the technique provides a 2-D plot giving average relative conductivity up to the depth of resolution of the equipment. A 3rd dimension is added when the procedure is repeated at particular time intervals. 2.7 The instrument measures relative conductivity of a given area of the measured surface relative to a chosen zero point. Absolute values cannot be determined. Hence, the instrument is incapable of differentiating between a dry structure with wet areas and a wet structure with saline areas. Advantages 2.8 Since no coupling or contact with the surface of the structure is required, the surface of the structure remains unmarked. As a result of the portability of the instrument and the continuous emission and receptivity of electromagnetic fields, structures can be tested rapidly and without disruption of other activities. 2.9 An important feature of the conductivity methods is that the measurements can be repeated at intervals and changes to the internal fabric of the bridge can be monitored with time. This is particularly important in the monitoring of the long-term effects of remedial operations, on the internal condition of the structure. Limitations 2.10 Tests should be conducted under ambient conditions. However, heavy rain or other such adverse conditions, such as condensation, may cause erroneous results. Heavy water ingression can significantly affect the results of a conductivity survey if carried out in horizontal mode. However in vertical mode, which has a lower surface sensitivity, the results are usually not affected by heavy rain conditions, as most of the water is shed along the surface of a structure as run-off Values of conductivity are usually recorded when direct current is employed for the measurements but it must be noted that the electrical properties of the sample may vary with the instrumentation frequency, the instrument coil separation, the frequency of investigation and the magnetic permeability of the various strata under examination The value of conductivity read on the instrument does not represent the conductivity at any particular depth, rather the value is a function of all the matter between the face of the instrument and the maximum depth of exploration The lower the frequency, the deeper the penetration but the poorer the resolution as amplitude decreases exponentially with depth The depth penetration is given in Table 3.1 along with the accuracy. Realistically, only a small conductivity meter can be used - such as the EM38. The EM31 is too large for most bridges. Equipment and Procedure 2.15 The system comprises a conductivity meter emitting continuously, and receiving electromagnetic fields through two coils. Typical commercial instruments for surveying masonry arch bridges are described in Chapter 3 Paragraph 8. The long axis of the instrument shall be held parallel to the ground at all times. AN.3.4-5

172 3. SURVEYING THE STRUCTURE OF MASONRY ARCHES Method of Application 3.1 The principle of the method is described in Chapter 2. The electromagnetic conductivity of a masonry structure is a function of the degree of the water saturation of the materials within it and their electrical properties. Electromagnetic fields are propagated into the structure and the variations in electromagnetic response of the materials within the structure can be measured. The 2D geometry of the variations can be determined, while assessments of the absolute degree of saturation are only possible over a number of repeated surveys over a protracted time period with calibrated moisture measurements. Accuracy 3.2 In-situ calibration is of great importance in the actual values of the readings obtained. It is recommended that, if conductivity surveys have to be repeated over a period of time, calibration settings should be recorded so that they can be exactly reproduced. Thus, measurements taken on different dates can be compared for structural condition monitoring. However, this is not possible for all instruments. For instance, the instrument set-up of the EM38 involves an electrical nulling of the n-phase component. The adjustment controls used in this process has no markings, and hence, the setting cannot be duplicated accurately. 3.3 Some data on depth penetration is given in Table 3.1 below: Intercoil Separatio n (m) Frequenc y (khz) Example of Commercia l System Dipole Mode Depth Penetration (m) Resolution Geonics EM38 Vertical Horizontal Geonics EM31 Vertical Horizontal Table 3.1 Depth Penetration ±5% at 30mS/m ±5% at 30mS/m ±5% at 30mS/m ±5% at 30mS/m Applications 3.4 Water ingress and moisture movement into structures is important in terms of structural durability. For example, if the road surface of a brick masonry arch bridge permits water entry then the soil fill above the arch barrel may become saturated. This can result in degradation of the mortar between the bricks giving rise to premature failure. Another example of water inclusion in masonry structures results from moisture capillary rise from the bridge foundation. The actual height of water rise on the inside of the wall may be required to be known this height is generally greater than that shown on the external wall surface. In the majority of cases, salt content is associated with water content in the structure. This phenomenon can also cause great damage to the structure and rapid decay of the masonry wall, and it is, therefore, a cause of concern. Thus, a non-invasive method of determining moisture movement behind or inside the masonry walls is of significant value. AN.3.4-6

173 3.5 Hence, conductivity measurements can be used to assess: moisture content in the masonry; height of moisture capillary rise. 3.6 The results can be used for a number of purposes: Advantages to identify changing moisture content profiles over a masonry bridge at a single point in time; to identify moisture content variations over time at identical locations on the structure; for detecting whether the section or element investigated is damp; using a tomographic analysis, the thickness of the stone walls could be estimated. This would depend upon achieving some clear distinction between material conductivities. 3.7 The low cost, in terms of time in the data collection phase and post-data collection analysis, makes conductivity surveying convenient for repetitive surveys. 3.8 The technique allows for high resolution in the conductivity variations detected, thereby improving the quality of the interpretation of the results. Some instrumentation is compact and light which makes it suitable for bridge applications. Other, like the EM31 or GEM300, due to their dimensions and weight, are not recommended for this type of investigation. The technique allows for lateral variations of conductivity to be measured accurately. Simple multi-layered conductivity surveys are also possible by raising the instrument above the material/structure surface. While this is relatively simple on extended vertical surfaces, it might become totally impractical in arch bridge testing, for instance under the soffit of arches. Limitations 3.9 The temperature dependence of the electrical conductivity of the electrolyte is almost entirely due to the temperature dependence of the viscosity of the liquid and a change in conductivity of 2.2% per degree may be expected. This phenomenon implies that for large seasonal changes of temperature, the conductivity over the normal range of ambient temperature may double. In all cases it is advised that the results are calibrated against some physical measurement on the structure, such as a core hole. This phenomenon should be noted when the 3rd dimension of a survey is time The close proximity of metallic objects will severely limit the validity of the data. Data are affected not only by metallic objects, but also by ferromagnetic material, and by sudden changes in magnetic susceptibility and viscosity. In the case of brick masonry arch bridges, engineering bricks will have a significant effect on results. Hence, different materials affect the EM38 in different ways: copper, with a relatively low magnetic susceptibility and high conductivity, will affect mainly the quadrature component. Magnetite, with a high magnetic susceptibility and low conductivity, will affect mainly the in-phase component (with an associated effect on the quadrature readings), while iron, with a high magnetic susceptibility and high conductivity will affect both. As both the in-phase and quadrature components are scalar components of one vector signal, large changes to one affect the other. Similarly, measurement of the quadrature phase relies on the assumption that the received signal has an exact 90 o phase lag. In cases of high magnetic susceptibility material, however, this is not necessarily the case and the situation is not clearly understood. AN.3.4-7

174 3.11 For this reason it is not advisable to commission this type of test on structures with evident presence of ties or other metallic objects on the surface. Conversely, the technique may identify cleverly hidden metal objects. Equipment and Procedure 3.12 Conductivity meters can be operated both in vertical and horizontal mode (orientation of the instrument relative to the tested surface). When the instrument is held perpendicular to the measured surface (termed vertical orientation), the measured signal is representative of the condition of the structure between 0.1m and 1.5m from the surface. In this orientation, peak sensitivity is at a depth of c.400mm. When the instrument is held parallel to the measured surface (termed horizontal orientation), the measured signal is representative of the condition of the structure between the surface and 1m depth. In this orientation, peak sensitivity is at the surface. Typical commercial instruments would be operating at approximately 15kHz for penetration depth range up to metres depending on whether the instrument is used in horizontal or vertical orientation. A lower operating frequency (around 10kHz) could be used for penetration depths up to 6m in vertical dipole mode and 3m in horizontal dipole mode. The coil separation distance, in this case varying between the two instruments from 1m to 3.66m, also affects the depth of penetration A digital data logger can be attached to continuously record the output from the meter. Figure 3.1 illustrates the use of equipment and data logger on a stone masonry wall - in the vertical position. Figure 3.2 illustrates the equipment being used on a brick masonry arch in the horizontal position. AN.3.4-8

175 Figure Conductivity instrument operating on a masonry structure - in the vertical position against the wall AN.3.4-9

176 Figure Horizontal orientation EM38 data being collected from the upstream (North West) elevation of the spandrel wall at 0.5m elevation 3.14 The choice of conductivity meter depends on the dimensions of the structure to be tested and on the depth of penetration desired. Quadrature-Phase and In-Phase readings are both to be collected or the surveyor can limit the readings to one of the two phases, depending on the aim of the test. By measuring the quadrature component, the Geonics EM38 instrument can directly measure the apparent conductivity of the ground down to a maximum depth of about 1.5m with an accuracy of approximately 5% The procedure will involve marking out the masonry wall in traverses either horizontally or vertically. If the meter is to be operated in automatic mode, readings will be collected continuously along the survey lines. If the meter is operated in manual mode, stations will have to be marked on the traverses, at regular intervals related to the intercoil spacing. Maximum recommended spacing between reading stations is equal to spacing between transmitter and receiver on the meter. A denser grid of reading points will give a better resolution in the final contour map. The lateral extent of the volume whose conductivity is sensed by the meter is approximately the same as the thickness depth Test locations are dictated by engineering objectives, however, an attempt should be made to measure the variation in material quality or condition throughout the largest possible area of the structure typically an area of 3m x3m minimum. From such a large map of the conductivity distribution in the sub-surface, it should be possible to identify areas with greater gradient in conductivity, denouncing anomalies The number of readings required is dependent upon: the accuracy and resolution required in the evaluation; the instrument used; the operating mode of use of the equipment. AN

177 Interpretation 3.18 For materials with conductivity between 1 and 100mS/m the electrical properties which control the current flow are relatively independent of frequency and the DC or low frequency conductivity measured with conventional resistivity and conductivity equipment will essentially be the same as that measured using low frequency (up to 300kHz) electromagnetic techniques When interpreting results on wing walls it should be considered that most soil and rock minerals forming building materials are insulators and conduction through the rock matrix only takes place when certain clay materials, native metals and graphite are present. The minerals in the sand and silt fractions of the soil behind the wing walls are electrically neutral and are generally excellent insulators. The electrical conductivity of the material is, thus, primarily controlled by the particle size, the amount of water present in the pores and by the conductivity of the pore fluid. The general trend is that conductivity will increase with reducing particle size, increasing moisture content and increasing salt content Measurements made on material as a function of the moisture content by weight, show a conductivity that increases approximately as the square of the moisture content The solutions of salts in pore water will substantially increase the material conductivity Unconsolidated materials at temperate ambient temperatures usually display a range of conductivity between 1 and 1000mS/m, whilst the conductivity of rock lies between 0.01mS/m and mS/m. The conductivity of masonry structures and walls made up of natural building materials can be expected to be in a range between 0 and 150mS/m Table 3.2 gives a broad indication of the conductivity of geological building materials but extreme caution must be exercised in employing these values for anything other than a rough guide. Material Type Conductivity range (ms/m) Sandstone masonry Conglomerates Sandstone Limestone Loose Sand Alluvium and Sands River Sand and Gravel 7 10 Clays Argrillites Top Soil Table 3.2 Typical Values of Conductivity for Geological and Building Materials 3.24 Data should be plotted so as to obtain the contour of the area investigated (Figure 3.3), or pseudo three- dimensional distribution of the conductivity (Figure 3.4). Plotting of conductivity values across sections of the structure is also a possibility (Figure 3.6). Commercial software is available for producing 2-D contour map plots and tomographic elaboration of the section investigated. AN

178 3.25 There are a number of ways of elaborating and presenting the data: data can be simply superimposed on the wall drawing or image, as per Figure 3.3 or plotted on to a CAD-CAM generated 3-D type plot of a structure: building or bridge as per Figure 3.4 and 3.5; another procedure is to use tomographic software to produce a horizontal slice Figure 3.6. Figure Conductivity distribution on the wall of a masonry bridge for measurement depths up to 1.5m - see Figure 3.4 for colour legend AN

179 Figure Conductivity distribution on the upstream wingwall and abutment wall - a close-up example and the colour coding is shown in Figure 3.5 Figure Conductivity distribution on abutment wall - pink scale shows high conductivity whilst turquoise shows low conductivity Figure Conductivity distribution of cross section, at level 2m AN

180 Example An example of a conductivity survey on a sandstone masonry structure is reported below for the case of a stone masonry arch bridge (Figure 3.7 a) and b)). a) White line marks level of radar tomography, reported elsewhere (Colla, 1997). b) Main dimensions and wing wall where sonic tests were performed. Figure North Middleton Bridge, downstream view AN

181 3.26 The meter used has intercoil spacing of 1m and provides a maximum depth of exploration of 1.5m in Vertical Dipole Mode (0.75m in Horizontal Mode) operating at a frequency of 14.6kHz The meter has been used for both the upstream and downstream sides of this 2-span bridge and on one abutment wall beneath the main vault. The measurement stations followed a grid marked on the walls, in an area well clear of any evident metallic objects (drains, reinforcing beams). For maximum accuracy and good spatial resolution, measurements have been overlapped by having reading stations every half metre, both in the horizontal and vertical direction of the wall surface. The lateral extent of the volume of structure whose conductivity is sensed, permits accurate measurement of small changes in conductivity, for example of the order of 5% or 10%. Contacting and non-contacting at 0.25, 0.5 and 0.75, 0.9m distances from the wall surface conductivity measurements were taken, to obtain data at different depths inside the structure see Figure 3.8 below. Data were collected in a digital data recorder and later transferred to a PC for elaboration and presentation in 2-D contour maps and section plots. a) Instrument lifting positions b) In contact in vertical position Figure In situ conductivity data collection 3.28 The values obtained are in a high and very wide range: conductivity readings registered were as high as 120mS/m. The highest values were recorded on the downstream side with an average of 60mS/m, and lowest on the abutment wall (average of 38mS/m), whilst the upstream side registered an average conductivity value of 40mS/m Such values indicate heterogeneity in soil filling in the abutment, variations in moisture content and salinity The results of the survey are given in Figures AN

182 Documentation and Reporting 3.31 The report on the electrical conductivity investigation should include the following additional information: location of the bridge plus relevant construction details; site testing procedures and any calibration checks; state whether vertical or horizontal mode is used; analysis and graphical presentation of the data; contoured plot of the data; give the results of any complementary tests; interpretation of the data; recommendations and conclusions Reference to background theory and examples of applications are listed in Chapter 5. AN

183 4. COMMISSIONING AND SPECIFICATION OF ELECTRICAL CONDUCTIVITY 4.1 The following requirements are specific to the use of Electrical Conductivity for the investigation of Masonry Arches and should be used in conjunction with those given in Advice Note 1 General Guidance of this series of Advice Notes. Information Supplied to Tenderers 4.2 The tender documents should require the following additional information to be provided: whether electrical conductivity is to be used initially to give indications as to the likely applicability for the use of GPR; whether the test is solely to add additional information to a structural survey. Information Required of Tenderers 4.3 Format of data presentation. Electrical Conductivity Report 4.4 The Specification should require the following additional information to be included in the report on the electrical conductivity investigation: whether vertical or horizontal mode is used; the results of any complementary tests; expected accuracy of results; moisture contents otherwise tested; secondary inferred information. AN

184 5. SOURCES OF FURTHER INFORMATION Binda, L., Colla, C. & Forde, M.C. Identification of moisture capillarity in masonry using digital impulse radar, J. Construction & Building Materials, 1994, 8, No. 2, Colla, C. (1997), NDT of masonry arch bridges, PhD Thesis, The University of Edinburgh, Dept Civil and Environmental Engineering, Edinburgh, 242 pp. Colla, C., Das, P.C., McCann, D.M. & Forde, M.C. (1995) Investigation of stone masonry bridges using sonics, electromagnetics & impulse radar, Proc. Int. Symp. Non-Destructive Testing in Civil Engineering (NDT-CE), BAM, Berlin, Germany, September 1995, Vol 1, Colla C., Das, P.C., McCann, D.M. & Forde, M.C. (1997), Sonic, electromagnetic and impulse radar investigation of stone masonry bridges, NDT & E International, Vol. 30, No. 4, Colla, C., Forde, M.C., McCann, D.M. & Das P.C. (1995) Investigations of masonry arch bridges using non-contacting NDT, Proc. 6 th Int. Conf. Structural Faults & Repair-95, London, July 1995, Vol. 1, Engineering Technics Press, Colla, C., McCann, D.M., Das, P.C. & Forde, M.C. (1996) Non-contact NDE of masonry structures and bridges, Proc 3rd Conf on Nondestructiive Evaluation of Civil Structures and Materials, University of Colorado at Boulder, 8-11 Sept 1996, Colla, C., McCann, D., Das, P. & Forde, M.C. (1997), Investigation of a stone masonry bridge using electromagnetics, Evaluation and strengthening of existing masonry structures, (Binda, L., Modena, C., eds.), RILEM, p Culley, R.W., Jagodits, F.L. & Middleton, R.S. (1975), E-phase system for detection of buried granular deposits, Symposium on Modern Innovations in Subsurface Explorations, 54 th Annual Meeting of Transportation Research Board. Geonics Ltd. (1992) Geonics EM38 Ground Conductivity Meter Operating Manual, Geonics Ltd: Ontario, Canada. Heiland, C.A. (1968), Geophysical exploration, New York, Hafner Publishing Co. Keller, G.V. & Frischknecht, F.C. (1966), Electrical methods in geophysical prospecting, Ch. 1. Pergamon Press, N.Y. McNeill, J.D. (1980), Electrical conductivity of soils and rocks, Ontario, Geonics Limited, Technical Note TN-5, 22 pp. McNeil, J.D. (1980), Electromagnetic terrain conductivity measurement at low induction numbers, Ontario, Geonics Limited, Technical Note TN-6, 15 pp. Olhoeft, G.R. (1975), Electrical properties of rocks, The physics and chemistry of rocks and minerals, J. Wiley and Sons, N.Y., p Olhoeft, G.R. (1977), Electrical properties of natural clay permafrost, Can J. Earth Science, vol. 15, p Reynolds, J.M. (1997) An Introduction to Applied and Environmental Geophysics, Wiley: Chichester, UK. Smith-Rose, R.L. (1934), Electrical measurements on soil with alternating current, Proc. IEE, 75, AN

185 Tamas, F. (1982), Electrical conductivity of cement pastes, Cement and Concrete Research, 12: R.C. de Vekey ed (2001) Rilem Recommendations of Non Destructive Tests For Masonry, MS.D.3: Radar Investigation of Masonry, pp ; MS.D.4: Measurement of Local Dynamic Behaviour for Masonry, pp ; MS-D.8: Electrical Conductivity Investigation of Masonry, pp ; Materials and Structures, 34, 237, pp , April) (PD143/99). Ward, S.H. & Fraser, D.C. (1967), Conduction of electricity in rocks, Ch. 2. Mining Geophysics, Soc. of Exploration Geophysicists, Tulsa, Oklahoma, vol. 2. Young, M.E. (2002). Using electrical conductivity for assessing chemical residues in stone. In: Proceedings of Swapnet Understanding and Managing Stone Decay. Richard Prikryl & Heather A. Viles (eds.). Prague, Czech Republic, 7th - 11th May ISBN pp AN

186 ADVICE NOTE 3.5 GROUND PENETRATING RADAR (GPR) Contents Chapter 1. Introduction 2. Details of Technique 3. Commissioning and Specification of GPR 4. Sources of Further Information AN.3.5-1

187 1. INTRODUCTION 1.1 This Advice Note gives guidance on the use of Ground Penetrating Radar (GPR), for the investigation of: (a) Masonry Arch Bridges, and (b) Post-tensioned Concrete Bridges with plastic tendon ducts. AN.3.5-2

188 2. DETAILS OF THE TECHNIQUE Principle of Method 2.1 The principle behind the test is one of applying high frequency electromagnetic impulses to the structure through the use of antennae to probe the subsurface. Radar antennae for structural engineering applications operate at centre frequencies between 100 MHz and 1.5 GHz or higher. (Note: 1,000MHz = 1GHz) 2.2 The wave produced may be travelling through a multi-layer system: for example: fill/macadam, brick or stone masonry, mortar and possibly a defective zone; or concrete, voids and so on. The electromagnetic pulse will be partially transmitted and partially reflected at each change of interface represented by a change in the dielectric properties of the structure material. By recording the energy reflected from (or, alternatively, transmitted through) different interfaces, a representation of the subsurface may be built up. Surveys can consequently be performed in reflection or transmission mode. The data recorded are stored in the time domain and may be displayed either on a computer screen, printed out on a chart recorder or stored for future analysis on a computer. Since the energy radiated by the antenna is a divergent beam, reflections from targets (any layer or anomaly) may be recorded even if the antenna is not positioned directly above them, with the consequence that the true shape of the target appears modified and distorted on the radar plot. Hyperbolae (arches) displayed on this image indicate the position and depth of discrete objects in the survey area. Accuracy 2.3 The accuracy of a radar wave is complex and a matter of some debate. However, in simple terms, it should be related to the centre frequency of the radar wave, electrical permittivity (or dielectric constant, ε r ), and wavelength (λ) of the material through which the wave is propagating. 2.4 Recent examples of work giving further information on this topic include: (Colla, Forde, Das, McCann, & Batchelor, 1997), (McCann & Forde, 2001), (Colombo, Giannopoulos & Forde, 2002). 2.5 Further discussion on other aspects of accuracy related to clutter and re-bar sizes is given elsewhere: (ACI 228.2R-98), (Bungey, Millard & Shaw, 1994), (Concrete Society, 1997), (Padaratz & Forde, 1995 a&b). 2.6 A simplified approximation is given in Table 2.1 below. This is adequate as a basis for the first choice of a radar antenna for a field survey. Resolution is taken as λ/2 and the depth of the first detectable target, Z min, is taken as λ/3. AN.3.5-3

189 Material ε r Frequenc y in air (MHz) Frequency in material (MHz) Velocit y (cm/ns) Wavelengt h (cm) Resolutio n (cm) Z min (cm) Penetratio n (cm) Masonry 6 1,500 1, Very low Concrete 10 1,500 1, Very low Masonry Low Concrete Low Masonry Medium Concrete Medium Masonry High Concrete High Where ε r = dielectric constant (real) Applications Table 2.1 GPR Propagation Through Concrete and Masonry 2.7 The potential of radar for monitoring purposes is clear. The technique can be employed both for structural monitoring in time and for checking the outcome of repair interventions. Moreover, its feature of providing a 2-D plot of the section or element investigated makes this technique appropriate for detecting hidden features and construction characteristics (which may provide a reserve of strength when assessing masonry arch bridges). A more sophisticated computer reconstruction can provide 3-D images. 2.8 The dielectric constant of the fill is related to the moisture content. GPR measurements designed to find the dielectric constant can, therefore, provide an indication of the degree of wetness of the material being investigated which may be important in relation to the durability of brick arch rings. Equipment 2.9 The impulse radar system comprises a number of components Figure 2.1 a radar pulse generator of varying pulse repetition rate; appropriate centre frequency antenna; appropriate data recording system either digital on tape, hard disk or analogue on paper; data display. The specific settings for the radar system should be undertaken in accordance with the manufacturer s handbooks. AN.3.5-4

190 Figure Basic Principles and Components of a Radar System with a Monostatic Antenna Operating in Reflection Mode 2.10 The electromagnetic impulse is emitted as a complex partial sine wave via an antenna of a certain centre frequency. For many engineering applications, antennae available will operate at centre frequency in the range of 100 MHz to 1.5 GHz or higher. For structures of considerable dimensions and in less favourable material conditions, lower frequency antennae (in the range of 100 to 500 MHz centre frequency) may be useful. For example, a 1 or 1.5 GHz antenna may be used for thin panel reinforced concrete or brick masonry when attempting to identify reinforcing bars. For double skin masonry it may be necessary to use a higher power 900 MHz antenna. For the investigation of stone masonry structures it may be necessary to use lower frequency as low as 100 MHz. Antenna devices may contain both a transmitting and a receiving bow-tie in the antenna casing. When a reflection survey is carried out in monostatic mode, in between pulsing, the antenna switches from transmitting to receiving mode The control unit records and merges the radar data to create a 2-D image. The waveforms are digitally sampled and analysed in terms of the magnitude of their deflection from the zero line. The minimum to maximum amplitude is generally split into 16 equal sections and each section is assigned a colour or a grey tone (Figure 2.2a). Waveforms are then placed adjacent to each other and plotted as depth (or two-way travel time, in reflection mode) on the vertical axis, against horizontal distance moved by the antenna on the horizontal axis. The magnitude of deflection from the zero line is shown as a point of colour. A complete profile is known as a line scan image. Alternatively, the data can be displayed as a time domain plot or wiggle plot, with a discrete number of waveforms sampled by the radar built-in oscilloscope (Figure 2.2b). On radargrams, the unit for time is nanoseconds and for distance is metres. AN.3.5-5

191 a) Line Scan Image in Grey Tones b) Wiggle Plot Figure Radar Data Display 2.12 Test locations are dictated by engineering objectives, however, an attempt should be made to measure the variation in material quality or condition throughout the structure. It is normal to test the structure in cross sections either horizontally or vertically. Thus, a series of traverses (horizontal or vertical) should be marked out for investigation usually on a 10cm to 1 metre or similar grid. (Prior to this, a quick survey of material uniformity may indicate that certain areas are of greater interest than others. Generally, a significant length of traverse, at least 5 metres, will be required in order to obtain an adequate basis for comparison of the data). The dimensions should be taken with considerable accuracy in order to enable subsequent mapping to be undertaken The system illustrated in Figure 2.1 shows a monostatic mode of operation (when the transmitting and receiving bow-ties are contained in the same antenna box); bistatic mode may be used when two separate transmitting and receiving antennae are used Usually antennae are dragged manually along the surface to be scanned. A survey wheel can be attached to the antenna to mark the radar plots at regular intervals of distance travelled or to record a constant number of scans per unit distance travelled The choice of antenna centre frequency depends upon the depth of penetration and the type of material to be tested and the resolution required - see Table 2.1 above for guidance. Depending on these factors and the survey requirements, the radar survey can be performed in reflection mode or transmission mode: 2.16 Reflection mode (Figure 2.1) is performed with one or more antennae located on the same side of the structure Transmission mode (Figure 2.3) is performed with the use of two or more antennae located on opposite sides of the structure. These transmission forms of radar surveying form the basis of radar tomography. AN.3.5-6

192 Figure Section Coverage Obtainable with a Single Transmission Scan: a) transmitting antenna (T) scans while receiver (R) is stationary; b) both antennae move in parallel; c) stationary transmitter and mobile receiver When the nature or the dielectric properties (permittivity and conductivity) of the media investigated are known, the surveys are aimed at calculating the depth of any target present. This is achieved via the use of the following simple velocity relation valid for dielectric materials and the signal travel time read on the radargram (Fig.2.4). Figure Representation of Arrival Time: Shorter Time in Correspondence of the Air Void Radar wave velocity, where, c = speed of light ε r = relative dielectric constant of material some typical values are given in Table 2.2 AN.3.5-7

193 Material ε r σ (ms/m) Material ε r σ (ms/m) Air 1 0 Dry clay Metal (iron) Saturated clay* Fresh water 81 1 Rock 4-10 Sea water 81 4 x 10 3 Dry granite Dry sand Wet granite 7 1 Saturated sand* Limestone Soil (dry) 2-6 Wet sandstone 6 Soil (wet) 5-15 Dry concrete 6 1 Clays Saturated concrete * Fresh water Table Dielectric Constant and Conductivity of Common Materials GPR Method Contractor s Equipment Spec. Advantage Relative Advantage Disadvantage Cost 1. Analogue radar system Low cost Visual interpretation of raw data only. Not recommended. Low cost survey 2. Digital: single channel; mono-static bow tie antenna Facility for both visual interpretation and post-processing More flexible interpretation than analogue method 1 Resolution and penetration into the structure is limited Medium cost 3. Digital: single channel bi-static Facility for both visual interpretation and post-processing Better resolution than monostatic, method 2 Medium cost 4. Digital: multichannel Facility for both visual interpretation and post-processing Potential for higher resolution than single channel bi-static, method 3 Higher cost 5. Post-processing software Facility for cleaning-up data; more effective on bistatic; and even more effective on multichannel May give better interpretation than methods 2-4 without signal processing Requires extra processing on a PC; requires more technical expertise Table Application Table of GPR Methods Extra over cost 2.19 Alternatively, when the antenna-to-target distance (or antennae separation) is known, the survey is aimed at velocity sounding. From the calculation of the signal velocity, the average dielectric constant of the material may be computed and the nature and condition of the material can be evaluated At an operational level the radar system should ideally be used in conjunction with a survey wheel attached to the antenna to improve the locational accuracy of the survey Table 2.3 lists some of the radar systems available. Tests will be conducted under ambient site conditions, within the operating temperature specification of the radar system manufacturer. AN.3.5-8

194 Limitations 2.22 In general, high frequency antennae give good spatial resolution but only shallow penetration as the signal becomes rapidly attenuated and may suffer from clutter (Padaratz & Forde, 1995 a&b). Lower frequency antennae are better suited to penetrate deeper as they emit more powerful signals, but their longer wavelength is a detriment to spatial resolution. Also, targets of small dimensions or small thickness can be missed see Table 2.1 for guidance The electromagnetic radar signal will transmit through non-conductive media (or insulators) at high speed, but will have very limited or zero penetration through conductors. These are materials characterised by high conductivity either because of water and/or salt content or because of their intrinsic characteristics. For example, radar will not penetrate through metals, nor will it penetrate through sea water (3.5% salt by weight concentration). The material conductivity is the main factor of signal attenuation and this will affect the exploration depth of the survey. Studies of attenuation and reflected signal amplitude measurements show that little penetration will be achieved above 0.25% salt concentration in water, with loss of 50% of the signal energy (Colla, 1997) A further limitation to the use of GPR would apply to surveys on brick masonry arch bridges, where engineering brick was used and they might have a high conductivity thus attenuating the GPR signal substantially From the above paragraph, it will be clear to the design engineer that there is little point in specifying a radar survey when the water within the structure to be surveyed is saline. The key to making the correct decision is the execution of an electrical conductivity survey before a radar survey is undertaken since the measured conductivity values can be directly related to the salinity of the water within the structure. When the salt water content is greater than 0.05%, the survey will be ineffective since even at relatively low saline contents, penetration will only be gained with a low frequency antenna resulting in a reduction in the resolution of the survey Clutter refers to signals returned from small scale heterogeneity in the materials investigated (cracks and joints, mortar joints, delaminations, small voids) and the phenomenon increases the higher the antenna frequency used. For the radar signal to penetrate through a heterogeneous material, the wavelength should be large compared with the typical heterogeneity or clutter dimension (5 to 10 times longer). For example, typical wavelengths in masonry may be between 100 and 1,500mm according to antenna frequency (see Table 2.1 above). Note that the target sought should also be considerably larger than the clutter dimension. For example, masonry in dry conditions is a dielectric material, and radar signals will propagate well through it, but the problem with older lower quality masonry is that in electrical terms it is lossy or may cause the signal to be rapidly attenuated and scattered The work should not be carried out in heavy rain or other adverse conditions as these will cause severely erroneous results. Heavy water ingression will significantly affect the results of a radar survey. Testing on a wet surface will result in the signal being reflected from the water and little penetration of the surface. Advantages 2.28 The principal reason for using ground penetrating radar (GPR) is that it is a non-invasive technique, with a potential for deep penetration at low resolutions and a higher resolution at shallow penetrations, depending upon the frequency of the antenna chosen and the nature and conditions of the materials investigated. No coupling is required. Radar will penetrate through air and structures presenting air filled compartments, but will not penetrate through metal. The presence of any metal in the masonry such as pipes, drains and reinforcement will be easily identifiable. AN.3.5-9

195 2.29 Due to the rapid pulse repetition rate of radar (typically 50 khz) the structure can be tested at rapid scan speed. Interpretation 2.30 The data may be expressed as the time to reflection from a target ; or the depth of penetration of the radar; or velocity; or average dielectric constant. The data may be displayed in three different formats: grey scale line-scan: this is a common form of display (Figure 2.2a); colour line-scan: this can be used to highlight certain features; wiggle-plot and single waveforms: these can be used to enable detailed examination of a single feature and precise calculations of target depth (Figures 2.2b and 2.5); the sending pulse can be seen at -1.3 ns and the received pulse (or reflection) at 3.63 ns. Figure Measurements of 2-way Travel Time from Single Wiggle Plot Reflection Recording AN

196 2.31 Interpretation of radar data is an objective dependent activity and can also be subjective to the individual carrying out the interpretation. Aspects to be considered for a valuable interpretation of data include: a clear understanding of the survey aims; an understanding of the possible configuration of the structure investigated; systematic organisation of the data in sets to be correlated with structural drawings; an estimation of the material properties; an estimation of the signal propagation characteristics (velocity and attenuation, achievable depth of exploration) as a function of the material properties; possible radar response expected; plan calibration drilling or follow up control work If one is looking for layer changes in masonry or concrete structures, then a grey scale or suitably chosen set of colour scales may be appropriate in identifying the time to reflection. This can be converted to depth if the dielectric constant is known. Alternatively, the average dielectric constant can be calculated if the depth is known. If a cross-sectional image of the structure is required, then complex cross-shooting of the radar impulses will be required see the survey of a masonry arch bridge under Example below The resolution of the signal is dependent on its wavelength (λ) and is generally estimated to be λ/2. The depth of the first discernible target is λ/3. However, in practice the frequency decreases due to antenna coupling effects and attenuation of the higher frequency components of the signal and the resolution suffers from these effects. See Table 2.1 for an approximate guide to the resolution of most commonly used antennae on concrete and stone masonry An approximate guide to select antenna frequency according to depth, based on practical experience, is given in Table 2.4. The antenna frequency (f) is closely linked to signal velocity (v) and wavelength (λ): v = λf. The use of different antenna frequencies may lead to small differences in the velocity value calculated for a specific material. Depth (m) Centre Frequency (MHz) Table Antenna Frequency as a Function of Exploration Depth in Reflection Mode AN

197 Internal Dimensions and Shape 2.35 GPR results on concrete bridges and masonry arch bridges indicate that if there is some reference point which ties in the GPR record with the wall thickness, then the data may be used to infer the wall s thickness at other points. However, using the GPR data alone it is not possible to determine the wall s thickness with a high degree of certainty. This is due to the often disrupted signal from the backwall combined with reflections from internal irregularities which leads to reflections from the back face of the wall which are not entirely distinct from other signals in the record As mentioned above it is necessary to know the velocity with which the signals travel through a medium to determine thickness. For GPR the dielectric constant is the parameter controlling the velocity as indicated above. Table 2.2 gives values for materials used in the construction of bridges. However, it is best to determine the dielectric constant of the materials on site and for this to be carried out a section of the wall where the thickness is known is required and where the GPR reflections may be picked up The most suitable antenna for the inspection of voids behind the spandrel walls and arch barrel depends on the thickness of the wall. For walls expected to be around 1m or less, the 500MHz is likely to be most suitable. Penetration deeper into the fill material would require a lower frequency antenna. Due to the variability of the signals recorded along some walls it is expected that small voids (around 0.1m wide, though dependent on antenna used) may not be detected amongst competing reflections. From the road surface antennae of 500MHz and below are suitable A high frequency, high resolution antenna, 900MHz or higher, is the most suitable transceiver for detecting fine cracks near parallel to the surface If voids are present in the fill material behind the wall, and within range of the GPR system, reflections which show the opposite polarity relative to the surface reflection should be looked for. This is due to the dielectric constant of air being less than the dielectric constant of construction materials. However, this is not a guarantee that voids are present, as some other interface between a higher and lower dielectric material may be encountered, though air does present the greatest possible contrast apart from the presence of water or metal conductors Cracks and voids may be investigated using GPR. A near vertical disruption to the GPR record should be looked for and may identify cracks propagating through the structure normal to the surface. Cracks more parallel to the surface may be determined by a reflection opposite in polarity to the surface reflection. For example, the transition from masonry or concrete to an air filled horizontal crack would result in a phase change in the reflection signal as the value of the dielectric constant of air is 1, compared to 4 to 12 for stone or concrete see equation for the reflection coefficient, R, below: 2.41 The ability to identify cracks depends on the uniformity of the wall s construction; the more variable and irregular the construction type the more complex the GPR image (i.e. rich in reflections) which reduces the success in distinguishing defects such as cracks. AN

198 Example of GPR Testing of Masonry Bridges Masonry Arch Bridge Wing Walls and Abutments 2.42 A transmission radar survey was carried out on a twin arch masonry bridge by using two 100MHz antennae positioned on upstream and downstream sides of the wing walls of a stone masonry bridge. The transmitter, on the downstream east wing wall was moved along the white dashed line shown in Figure 2.6a, whilst the receiver was stationary, at the same level, on the upstream wing wall. The survey was repeated a number of times with the same movement procedure for the transmitter, whilst the receiver moved along in 1m steps to obtain an adequate coverage of the bridge section. Due to the slope of the river bank on the upstream side, movement of the receiver was limited. a) Downstream View. White dashed line marks position of transmitting antenna and direction of scan. b) Horizontal section showing coverage of radar tomography AN

199 Figure Middleton North Bridge 2.43 This antennae configuration, maintaining one antenna stationary whilst the other is moved, offers the advantage that more accurate readings can be recorded. In fact, only the movement of one antenna needs to be controlled and its position can be monitored accurately by marking the radar plots at regular intervals of distance travelled by the transmitter or by attaching a survey wheel to the moving antenna. In the case of Figure 2.6b, both antennae need to be dragged at exactly the same velocity to maintain them in a constant facing position so that accurate measurements of the travel time can be recorded The radar plots obtained are of the kind shown in Figures 2.7 and 2.8 where the arrival travel times are recorded at the receiving antenna location (upstream). This kind of time domain section through the structure forms the basis of radar tomography where the arrival times in each waveform or the signal amplitude are used as input parameters in the tomographic inversion, software such as MIGRATOM (Jackson & Tweeton, 1994), to plot maps of signal velocity or attenuation. Figure Radar tomographic plot obtained with 100MHz antennae. Rx marks position of the receiver on upstream side, whilst transmitter moves along downstream side Figure Radar tomographic plot obtained with 100MHz antennae. Rx moved further along on upstream wing wall (300 ns range) AN

200 2.45 The first finding from this radar survey is a generally low velocity of the electromagnetic (EM) signal which indicates the values of dielectric constant within the bridge to be significantly different from the values expected for construction or fill materials see Table 2.2 above. The average value for the dielectric constant ε r for the composite bridge construction (masonry/soil fill/masonry) on this site was computed to be approximately 56. This value is well above reference values published in literature and could be explained by a high moisture content in the fill, due to possible moisturedrainage problems in the bridge Consider now Figure 2.7 (a typical grey scale line scan from a radar survey of the kind depicted in Figure 2.2a) and the position of the receiver Rx in relation to the transmitter on the downstream side. As the EM waves travel in straight lines, the shortest ray paths and corresponding shortest travel times are to be expected when the transmitter reaches the position just opposite the receiver Rx assuming that the material within the bridge is homogeneous. Similarly, the received signal would show the maximum amplitude at this location (its attenuation being minimum there) whilst at longer transmitter/receiver distances the signal would show greater attenuation. In Figure 2.7 this is true for the received signals to the right of the Rx position, but not to its left hand side. In Figure 2.8 the situation is even more complex: at a transmitter position between 0 and 1m on the downstream side and a receiver position of 3.5m on the upstream wall, the shortest travel time registered on the scan is actually the longest distance travelled. Also, the attenuation of the signal is greater, compared with the case of Figure It can be postulated that the phenomena can be attributed to uneven dielectric constant values across the bridge section, with materials characterised by lower dielectric constants between 0 and 2.5m on the wing wall than between 2.5 and 5m. The attenuation of the received signal seems to follow the same trend seen for the velocity. Since it is particularly noticeable when the receiver is at location 3.5m, it can be deduced that materials on the right hand side of the wingwall are characterised by higher dielectric constants and also by higher conductivity values The above assumptions, based upon the simplified equation for EM wave velocity, need to be reconsidered in the light of the general equation for EM wave velocity, given that a material characterised by high conductivity and tested at low frequency may behave as a conductive material: [ ( )] where ; c = speed of light ε r = dielectric constant tanδ = loss tangent. (This is influenced by conductivity, frequency and dielectric constant) 2.49 It should be noted that the theory of Electromagnetic propagation and terms such as loss tangent are highly complex and commonly applied in simplified form as described in the literature (Padaratz & Forde, 1995; Daniels, 1996) The above relationship was plotted graphically in a parametric analysis linking material conductivity with antenna frequency to obtain signal velocity. In the case of this project, where 100MHz antennae were used on highly conductive media, the study shows that the signal velocity will decrease significantly with any increase in conductivity, as will the amplitude of the signal due to increased attenuation. Consequently, it is possible that the material dielectric constant has not actually changed rather the reduction in velocity results from the increase in conductivity. AN

201 Masonry Arch Barrel 2.51 Figure 2.9 shows the result of scanning the intrados of the arch looking upwards into the bridge. The objective was to trace out the arch thickness and to look for any structural faults. The upper part of the figure is a classically line-scan whilst the lower part is a single wiggle plot (rotated through 90 o for clarity) showing the time history record at the crown of the arch. A reflection which is parallel to the curved arch does appear on the GPR data, however, the two-way arrival time of this reflection is 1.6 ns which is quicker than expected based on an estimate of the dielectric constant, ε r, of 5 for the brick (determined from another GPR section and the arch thickness found by a previous trench). It is, therefore, shown that there are cases where a single GPR data section is insufficient to resolve structural detail. Figure GPR Profile of Arch Barrel (900 Mhz, 2 channels, 12 ns range) Following the Curve of the Intrados Masonry Arch Barrel and Fill 2.52 Figure 2.10 shows the data collected on scanning over a small masonry arch bridge of span 3.7m. The distance between road surface and soffit was 0.7m; around 0.3m of road pavement and fill above an arch 0.4m thick. This dataset is just able to make out the extrados and intrados of the arch which allowed dielectric constants to be estimated for the fill (40) and brick arch (5). The value of 40 for the fill is high and indicates that the fill is wet. The data section also indicates that the fill material is not as layered as the surrounding original soil and that the high water content is resulting in weak returned signals due to the increased conductivity resulting in the shape of the arch being attenuated. AN

202 Figure MHz GPR Profile (40 ns range), Arrow Marks the Arch Intrados. 1) Original Soil; 2) Fill, Along the Road Surface Over a Masonry Arch Bridge Stone Parapet and Spandrel Wall 2.53 Measurements were carried out with a 900MHz frequency antenna in reflection mode to investigate a stone parapet and spandrel wall thickness in a single span masonry arch bridge. A vertical scan was performed on the upstream side of the bridge, starting from the top of the parapet wall and moving downwards. (Figure 2.11). Data was analysed to identify if the spandrel wall had greater thickness toward its base and stepped rear face. Figure 2.11 Kilbucho Bridge: upstream elevation (dotted vertical line marks position and direction of scanning) 2.54 In the parapet wall the thickness was known and, from the measured 2-way travel time of 4.93 ns, a velocity of m/ns was calculated; this corresponds to a dielectric constant value of approximately ε r = 5.7. At the top of the spandrel wall a small increase in travel time was registered (t.t. = 5.88 ns) and attributed to a higher moisture content. The velocity and dielectric constant were calculated in a similar manner for the spandrel wall: v = m/ns and ε r = 8.1. Although the radar plot was not clear enough to allow calculation of the thickness of the wall, an increased thickness of the spandrel wall is registered in its lower position. AN

203 Stone Parapet and Spandrel Wall 2.55 The GPR section of Figure 2.12 shows the result of scanning a 900MHz antenna down the parapet and spandrel wall of a masonry arch bridge. From measuring the sandstone parapet thickness and recording GPR reflection data the dielectric constant was estimated. Using knowledge of the dielectric constant and by following the backface parapet reflection to the reflection observed behind the spandrel wall, the thickness of the spandrel wall may be determined. In this case the spandrel wall is estimated to be of the order of 0.45m thick. It is evident that the reflection from the spandrel backface is weaker, in some parts it is not present, and that it is of the opposite polarity to that of the parapet backface. This is due to the lower and opposite contrast in dielectric constant between the wall and air and the wall and fill - see paragraph 2.38 above N.B. The disturbance behind the parapet is a car driving across the bridge, metal being a very good reflector of GPR signals. AN

204 Figure MHz GPR Profile Down the Parapet and Spandrel Wall AN

205 GPR Testing of Post-tensioned Concrete Beams with Plastic Tendon Ducts 2.57 The testing procedures, which might be considered for these bridges, are listed in Table 2.5 below. In the table, the applicability to Metal and Plastic tendon ducts is indicated respectively. This part of this standard focuses on the use of GPR to detect voids in post tensioned beams with plastic ducts. Investigation Method Cost of Method Metal Ducts Plastic Ducts Effectiveness of Technique Visual Inspection Low No No Technique is ineffective as bridges rarely show distress before catastrophic failure. Load Test Relatively high No No Ineffective procedure and dangerous as the structure could fail before any meaningful deflection response is obtained. Stress/strain measurement Relatively high No No Generally ineffective as Cavell (1997) has shown that post-tensioned bridge strain variations due to loss of pre-stressing can be similar to variations resulting from temperature gradients throughout the year. Thus, this technique is not sensitive to the defects in posttensioned bridges. Impulse radar Intermediate No Yes Effective with non-metallic liners such as in the joints of segmental bridges and in the newer post- tensioned bridges. Radar will not penetrate post- tensioned metal ducts. Impact echo Intermediate Yes Maybe Potentially useful in identifying voiding in nonmetallic and metallic post-tensioned ducts. Essential to ensure that impact frequency is sufficiently high to identify the defect. Manual drilling of tendon duct with visual inspection using endoscope Intermediate Yes No Statistically limited and potentially dangerous if the tendons themselves are drilled. Advantage is that a direct physical observation can be made. Radiography High Yes Yes High powered radiographic techniques give good image of voiding but requires closure of the bridge and may not be used in urban areas due to the risk of radiation. Ultrasonic tomography Intermediate Yes Yes Promising technique that could identify voids by producing a 2-D or 3-D image of the beam cross-section. Table 2.5 Test Methods of P-T Concrete Bridges AN

206 Purpose of GPR in Detecting Voids in Plastic Tendon Ducts 2.58 This application of the radar technique (GPR) should be used to detect voids in grouted plastic posttensioned ducts, within concrete beams constructed as recommended in Concrete Society Technical Report TR47 (1996). The general principle of radar testing is given in under Principle of Method above. Earlier work on radar testing to identify voids in plastic ducted post- tensioned concrete beams was partially successful (Bungey et al 1997). More recent numerical and experimental work at the University of Edinburgh has shown increased promise in the application of the technique (Giannopoulos, Macintyre, Rodgers & Forde, 2002). Application and Methodology 2.59 Before starting the evaluation of the plastic ducts for voids, the location of the ducts should be established from as-built drawings or using digital impulse radar in the parallel configuration at cross- sections at appropriate intervals, see Figures 2.13 and The cross-section intervals would depend on the length of the beam and the curvature of the duct. The location of the test points must be accurately recorded The location of the duct to be investigated should then be marked up on the surface of the concrete. Figure Parallel Configuration of 1.5GHz Antenna Figure 2.14 Perpendicular Configuration of 900MHz 2.61 In order to investigate for voiding in the plastic tendon duct using GPR, the perpendicular orientation of the antenna should be used - Figures 2.14 and The choice of antenna centre frequency from 500 MHz to 1.5GHz will depend upon the depth of the duct and power required from the antenna. For example, it has been shown that the shallowest detectable target with a radar antenna is the wavelength/3 (Colombo, Giannopoulos & Forde, 2002) Table 2.6 below gives guidance on the choice of antenna: this table is based upon the standard relationship: Velocity = frequency x wavelength V = f x λ AN

207 2.64 In accordance with earlier research reported in Padaradz & Forde (1996), it has been assumed that the centre frequency of the impulse radar pulse through concrete is reduced by 30%, when the antenna is coupled to the concrete column 6 below. Column 1 Column 2 Column 3 Column 4 Column 5 Column 6 Column 7 Column 8 Column 9 Material ε r Frequency in air (MHz) Frequenc y in concrete (MHz) Velocity in concrete (cm/ns) Wavelength in concrete (cm) Resolution (cm) Zmin (cm) Penetration (cm) Concrete (very dry) Concrete (dry) Concrete (damp) Concrete (wet) Concrete (very dry) Concrete (dry) Concrete (damp) Concrete (wet) Concrete (very dry) Concrete (dry) Concrete (damp) Concrete (wet) 4 1,500 1, Very low (< 50) 6 1,500 1, Very low (< 50) 10 1,500 1, Very low (< 50) 20 1,500 1, Very low (< 50) Low (50-100) Low (50-100) Low (50-100) Low (50-100) Medium (> 100) Medium (> 100) Medium (> 100) Medium (> 100) Table 2.6 GPR Propagation Through Concrete where ε r = dielectric constant (real) Z min = minimum depth of detectable target Column 5 gives the velocity through concrete. (These figures are an extension of figures reported elsewhere, e.g. Concrete Society Technical Report No. 48, 1997.) It has been assumed that the concrete is of low conductivity and thus, the simplified velocity equation is valid: where c = speed of light AN

208 ε r = dielectric constant of the material. Column 7 assumes that resolution is based upon λ/2 (Martin, Hardy, Usmani, & Forde, 1998) due to the dispersive nature of concrete. Some might argue that a better resolution such as λ/4, may be achievable but this remains unproven. Column 8 is calculated upon the assumption that the shallowest possible detectable target is λ/3 from recent research in Italy (Colombo, Giannopoulos & Ford e, 2002). Column 9 gives general guidance on the likely penetration of radar antennae of different frequencies. Where values of depth penetration are given in cm units, these are only approximate and will vary with different manufacturers systems. Also note that as the moisture content of the concrete increases, so the signal attenuation increases Given the detailed guidance in Table 2.6, plus accompanying comments the testing organisation must give reasons for the choice of their specific GPR antenna Typical GPR results that might be expected in the field are shown below in Figures 2.15 and Note: there were no links included above the voided parts of the tendon duct in order to give maximum clarity of the voided section. However, in the field, it would be unreasonable to assume there would be no links above the voided parts of the duct. Figure 2.17 gives details of the construction of the beam. Figure Scan with 900MHz Antenna - Perpendicular Orientation AN

209 Figure Scan with 1.5GHz Antenna - Perpendicular Orientation AN

210 Figure Beam Details AN

211 Interpretation of Test Results 2.67 Interpretation of radar data is an objective dependent activity and can also be subjective to the individual doing the interpretation. Aspects to be considered for a meaningful interpretation of data include (Colla, Forde, Das, & McCann, 1997): a clear understanding of the survey aims; an understanding of the possible configuration of the structure investigated; systematic organisation of the data in sets to be correlated with structural drawings; an estimation of the material properties; an estimation of the signal propagation characteristics (velocity and attenuation, achievable depth of exploration) in function of the material properties; possible radar response expected; plan drilling or follow-up control work A grey scale or suitably chosen set of colour scales might be appropriate in order to identify the time to reflection. This can be converted to depth if the dielectric constant is known. Alternatively, the average dielectric constant can be calculated if the depth is known. (Colla, Das, McCann, & Forde, 1995; Colla, Das, McCann & Forde, 1997; Colla, Forde, Das, & McCann, 1997). If a cross-sectional image of the structure is required then complex cross-shooting of the radar impulses will be required. This might not be practicable on a bridge with an overlying deck. Example 2: Laboratory Experiments of GPR on P-T Concrete Beams with Plastic Ducts at the University of Edinburgh 2.69 From the work of Giannopoulos et al (2002) referred to above, it was established both numerically and experimentally that the optimum orientation for the radar antenna is a parallel direction for the location of voided plastic ducts. This is at 90 o to the normal method of operation in the field when testing beams. Defects 2.70 Defects were introduced into the beams to simulate air voids. These voids were introduced at the grouting stage. Experimental Design 2.71 Several test beams with known defects were designed in order to validate the results of numerical simulations of GPR. Beams with both metal and plastic ducts were constructed. Note that the metal ducts represent p-t beams prior to the 1992 Highways Agency moratorium on the use of p-t bridge beams (DTP Press Notice 260, 1992), whilst the plastic ducts represent post-moratorium p-t beams. A uniform cross-section of 400mm x 450mm and a length of 2000mm for all the beams was employed. For the results presented here one of these beams incorporating a plastic duct having 63mm external and 50mm internal diameters was used. A steel tendon with a diameter of 20mm was included at the centre of the duct. AN

212 2.72 In this concrete beam illustrated in Figure 2.18 the duct is completely grouted apart from an 800mm long section that is completely voided. The main aim of the experiment was to detect the voided section of the duct when scanning along the side of the beam with GPR. For this investigation the cover depth of the duct was 145mm. Results Figure Test Beam Design 2.73 Several GPR scans were obtained from the test beam using both the 900MHz and the 1.5GHz GSSI antennae. (GSSI = Geophysical Survey Systems Inc radar system) The aim of these scans was to determine if it was a realistic expectation to detect a completely voided duct which is the worst case scenario In Figure 2.19 the results from the 900MHz parallel scan are presented. (Parallel scanning is the normal site method of GPR testing.) Although the duct is clearly detected it is almost impossible to distinguish the voided section from the completely grouted one. The results obtained with the 1.5GHz antenna shown in Figure 2.20 exhibit a similar pattern. The orientations of the GPR antennae are parallel to the long axis of the duct. AN

213 Figure Scan with 900 MHz Antenna - Parallel Orientation Figure Scan with 1.5GHz Antenna - Parallel Orientation AN

214 2.75 When the same scans were repeated with the orientation of the GPR antennae aligned in such a way that they are perpendicular to the long axis of the duct, the results were significantly different from the parallel case. As is illustrated in Figures 2.21 and 2.22 for the 900MHz and the 1.5GHz antennae, the voided section is clearly identified. This was especially so in the case of the 900MHz antenna for which the cover depth does not present a significant obstacle to signal penetration as in the case of the 1.5GHz antenna. Figure Scan with 900MHz Antenna - Perpendicular Orientation Figure Scan with 1.5GHz Antenna - Perpendicular Orientation AN

215 2.76 The above results clearly suggest that a perpendicular orientation of the GPR antennae is most appropriate for the detection of voids in post-tensioned beams with plastic ducts. The use of both a 2- D and a 3- D GPR model provides an interpretation for these experimental findings. Conclusion from the Experimental Work at the University of Edinburgh 2.77 An investigation into the detection of voids in post-tensioned (p-t) concrete beams by ground penetrating radar (GPR) was undertaken. Both experimental and numerical modelling results suggest that the optimum orientation of the radar s antennae is perpendicular to the long axis of the ducts containing the post-tensioning tendons. Example 3: Numerical Modelling of GPR Experiments on P-T Concrete Beams with Plastic Ducts at the University of Edinburgh 2.78 The following detailed section is a numerical experiment of GPR waves through p-t concrete beams with plastic ducts. The conclusion that will be drawn from this section is: that the detection of voids in post-tensioned concrete beams with plastic ducts is easier with a perpendicular orientation of the GPR antennae compared to the parallel one that is often used in field work. Background to Numerical Modelling 2.79 The Finite difference time domain (FDTD) approach used to model GPR, is a numerical method that provides a solution to Maxwell s equations, expressed in differential form in the time domain. The method, originated by Kane Yee (1966), is based on the discretization of the partial derivatives in Maxwell s equations using central differencing. The resulting difference equations are used in a time marching iterative procedure to obtain the required solution. Since the appearance of Yee s original paper, the FDTD method has been widely used in the solution of a diverse range of electromagnetic field problems such as, radar cross-section estimations, EMP coupling to dielectric structures, antenna modelling, electromagnetic field penetration, propagation in plasma and biological applications. A detailed review of the FDTD method can be found in Taflove (1995). Because the FDTD formulation is entirely in the time domain it is particularly suited for solving transient problems such as GPR In all models presented here, all media are considered as linear, passive and non-magnetic. Furthermore, the velocity of propagation of the electromagnetic waves in concrete is assumed not to be frequency dispersive which is reasonable for frequencies above 500MHz at low conductivity. The electrical properties of concrete were set to ε r = 6 and σ = S/m where the parameters of grout are assumed to be very similar with the ones of concrete (same relative permittivity and σ = S/m). The plastic duct was modelled as a no-loss medium of relative permittivity ε r = In both 2D and 3D models the GPR antennae are modelled as idealised sources line source for the 2D case and Hertzian dipole for the 3D one and without any shielding. Although it is possible to include the actual antenna structure in the 3D model this would have resulted in an excessive requirement of computational power that is not commonly available to the testing engineer. All the numerical results have been obtained using the GPR FDTD simulator GprMax2D/ 3D ( developed by the Research Fellow (Giannopoulos) and running on a Personal Computer operating in Linux. AN

216 Results From Two-Dimensional GPR Models 2.82 In order to use a 2D model for GPR, some, specific assumptions are required: the geometry of the problem is invariant in the strike direction, which results in an infinitely long target in that direction. Therefore, only the cross-sectional area of the target is considered; the sources are also infinitely long and parallel to the strike direction; therefore, in a 2D model only (sources) antennae which are parallel to the axis of the ducts can be modelled; the geometry of the beam cross-section used in the 2D models is as in Figure 2.18; simulations have been performed with various grouting levels; it should be noted that the plastic walls of the duct were included in the models GPR responses from the 2D model were obtained for various centre frequencies when the GPR antennae were situated at the side of the beams and centred over the plastic duct In Figures 2.23, 2.24 and 2.25 the simulated GPR traces from the 2D FDTD model for GPR centre frequencies of 900MHz, 1.5GHz and 2GHz are presented. Figure Simulated 2D GPR Responses for 900MHz Centre Frequency AN

217 Figure Simulated 2D GPR Responses for a 1.5GHz Centre Frequency Figure Simulated 2D GPR Responses for a 2GHz Centre Frequency AN

218 2.85 The results from these 2D models suggest that it is quite difficult to distinguish the various degrees of voiding inside a plastic duct when the centre frequency is 900MHz. There are some differences for centre frequencies of 1.5GHz and more significant ones for 2GHz. Hence, the practical detection of voids using the commonly employed 900MHz antennae in parallel orientation is not trivial. From the 1.5GHz and 2GHz results it is clear that the 75% grouting looks similar to 100% and the 50% is closer to the response from a completely voided duct. These simple 2D models are in qualitative agreement with the experimental results for the parallel antenna orientation. Results from Three-Dimensional GPR Models Related to the Interpretation of the Experimental Results 2.86 An interpretation of the experimental GPR results when the polarisation of the antennae is perpendicular to the long axis of the duct can be obtained using a full 3D numerical model. The modelling procedure is identical to the 2D case with the obvious difference that the special assumptions for the 2D case are not applicable any more. For these 3D models, instead of using the actual cross-section of the test beam, a half-space concrete model was employed. The reason for such a simplification is that since in our numerical model the GPR antennae are not shielded, there will be contributions from the edges of the beam that may obscure the responses from the plastic duct. In real GPR surveys with shielded antennae, these edge effects are not very significant except when the antennae are in close proximity to such an edge. The fact that a half-space is used in the model does not affect the response from the plastic duct in any other way, since the cover depth has been maintained according to the beam design In Figure 2.26 simulated GPR traces from the completely voided and grouted duct are compared for centre frequencies of 900MHz and 1.5GHz. Note that the orientation of the GPR antennae is parallel with the long axis of the ducts. These results are very similar to the ones obtained from a 2D model. There is no significant difference between the responses from the completely voided and grouted ducts both are of similar magnitude and phase. Figure Simulated 3D GRP Traces for Parallel Orientation AN

219 2.88 In Figure 2.27 the same simulated traces for the case of perpendicular orientation are presented. It is evident that the response from the voided duct is much clearer and stronger than the response from the grouted duct. This verifies the experimental findings for the perpendicular case. Figure Simulated 3D GRP Traces for Perpendicular Orientation 2.89 An explanation for these results can be obtained using some simple 3D models. In Figures 2.28 and 2.29 we compare the simulated responses calculated for different antenna orientations for models that include: only the metal tendon without any ducting; only the voided duct with no tendon; and only the grouted duct with no tendon. For the case of the parallel orientation the most significant response is from case (a), the single tendon (no ducting). This was expected from such a long metal target. However, it is not such a useful case, because in such an investigation one assumes that the tendon is contained in a duct. Considering the results for the perpendicular orientation, the response from the void is more dominant than the one from the tendon. AN

220 2.90 For both the parallel and perpendicular orientations the strength of the maximum response is approximately the same however, it is due to the presence of different factors. In the case of the parallel orientation the strength and type of the response is determined by the presence of the tendon, whereas in the case of the perpendicular orientation the strength and type of the response is determined by the presence of the void. Since the tendon is always present there is not much variation on the recorded GPR responses to facilitate the detection of voids in the ducts. However, because voids hopefully do not form often inside ducts a perpendicular orientation for the GPR antennae is more likely to detect them. Note that the final response recorded by the GPR is not just the summation of the responses from the individual components (i.e. tendon, duct, grout or void) however, the above simple experiment demonstrates how important we expect to be their contribution to the total GPR response. Figure Simulated 3D GPR Traces for a) Empty Duct b) Empty Grouted Duct and c) Tendon only for the Parallel Antenna Orientation AN

221 Figure Simulated 3D GPR Traces for a) Empty Duct b) Empty Grouted Duct and c) Tendon only for the Perpendicular Antenna Orientation Conclusions 2.91 This experimental and numerical investigation on the detection of voids in post-tensioned concrete beams with plastic ducts leads to the conclusion that a perpendicular orientation of the GPR antennae is more beneficial than the parallel one that is normally used in field work. AN

222 3. COMMISSIONING AND SPECIFICATION OF GPR 3.1 The following requirements are specific to the use of GPR and should be used in conjunction with those given in Advice Note 1 - General Guidance of this series of Advice Notes. Information Supplied to Tenderers 3.2 The tender documents should require the following additional information to be provided. For Post-Tensioned Beams as-built drawings providing reinforcing details, duct types and sizes, profiles, locations and depth of cover giving the best available information; which ducts and which lengths of ducts are to be tested in detail; whether other ducts or intermediate lengths of the above ducts are to be tested at sample locations. Interval to be stated; confirmation that the tendon ducts are plastic; For Masonry whether electrical conductivity is to be used initially to determine whether the structural condition is suitable for the use of GPR. 3.3 The following additional deliverables from the testing organisations should be stated: diagrammatic type interpretation (e.g. CAD drawing). Information Required of Tenderers 3.4 The Tenderers should additionally be asked to state: trade name, model and frequency of both the main frame and the antennae, including date of most recent calibration; to specify the probable range and choice of antennae to be used, together with justifications, based on Table 2.1 and 2.6; to state the likely filtering to be used; to indicate the likely number of points to be recorded; to indicate the expected standard of interpretation of the results, e.g. voids detected or no voids detected together with a percentage confidence level. AN

223 How to Use the Results Additional Considerations For Post-tensioned Beams The following limitations should be considered. Dense reinforcement may affect the accuracy of the readings, or outer ducts may mask the readings from inner ducts. Drilling into ducts (with care not to damage the tendons) should be used to confirm the presence of voiding and to confirm absence of voiding at selected critical locations, as NDT techniques do not always provide unambiguous definitive results. Vacuum testing rather than GPR should be used to determine the volume of the voids in grouted ducts as a precursor to regrouting them. Guidance documents have been prepared on regrouting techniques (Ref Concrete Society Technical Report TR72 Durable bonded post-tensioned concrete structures 2010). GPR Report 3.5 The report should additionally state: whether used in monostatic or bi-static arrangement, the number and frequency of the operating antennae, the time range used and rate of data recording (number of scans per second, number of samples per scan, etc.), whether filtering and gain were applied and of which kind; use of survey wheel on the antenna and antenna orientation in relation of the direction of scan; calibration of radar equipment. Precise settings used on the instrumentation, as the systems have varying capabilities and data can be recorded in a number of ways; record the number and frequency of the operating antennae, the time range used and rate of data recording (number of scans per second, number of samples per scan, etc.), whether filtering and gain are applied and of which kind; use of survey wheel on the antenna and antenna orientation in relation to the direction of scanning; clear description of the methodologies of collection of the data: whether transmission or reflection mode survey has been carried out and the surface distance covered by transmitting and receiving antennae; test results compiled in the form of grey-scale plot accompanied by time window or depth scale; 2-D and 3-D interpretative model, other tabulation as appropriate. AN

224 4. SOURCES OF FURTHER INFORMATION ACI Technical Report 228.2R-98 (1998) Nondestructive Test Methods for Evaluation of Concrete in Structures, ACI, Farmington Hills, MI, USA, p. 62. Bungey, J.H., Millard S.G. & Shaw, M.R. (1997) Radar assessment of Post-tensioned concrete. Structural Faults + Repair-97, Engineering Technics Press, Vol 1, Cavell, D.G. (1997) Assessment of deteriorating post- tensioned concrete bridges, PhD thesis, University of Sheffield. Colla, C., (1997), NDT of masonry arch bridges, PhD Thesis, The University of Edinburgh, Dept. Civil and Environmental Engineering, Edinburgh, 242 pp. Colla, C., Das, P.C., McCann, D.M. & Forde, M.C. (1995) Investigation of stone masonry bridges using sonics, electromagnetics & impulse radar, Proc. Int. Symp. Non-Destructive Testing in Civil Engineering (NDT-CE), BAM, Berlin, Germany, September 1995, Vol. 1, Colla, C., Das, P.C., McCann, D.M. & Forde, M.C., (1997), Sonic, electromagnetic and impulse radar investigation of stone masonry bridges, NDT & E International, Vol. 30, No. 4, Colla, C., Forde, M.C., McCann, D.M. & Das, P.C., (1997), Laboratory modelling of radar propagation through masonry models, Proc. 4th Int. Conf. NDT-CE, University of Liverpool, 8-11 April 1997, Vol. 1, Colla, C., Forde, M.C., Das, P.C. & McCann, D.M., (1997), Radar imaging in composite masonry structures, Proc. 7th Int. Conf. Structural Faults + Repair-97, Edinburgh, 8-10 July 1997, Engineering Technics Press, Vol. 3, Colombo, S., Giannopoulos, A. & Forde, M.C. Accuracy of radar testing of masonry arch bridges, IABMAS 02, July 2002, UPC, Barcelona. Concrete Society Technical Report TR48 (1997), Guidance on Radar Testing of Concrete Structures, The Concrete Society, Slough, UK, pp 88. Concrete Society Technical Report TR72 (2010), Durable Bonded Post-Tensioned Concrete Structures, The Concrete Society, Slough, UK, pp 64. Daniels, D.J. (1996) Surface-Penetrating Radar, IEE, pp 296. DTP Press Notice No Published by DTP, London Giannopoulos, A., Macintyre, P., Rodgers, S. & Forde, M.C. (2002) GPR detection of voids in post- tensioned concrete bridge beams, 9 th Int Conf, GPR- 2002, Santa Barbara, CA, 29 th Apr - 2 May Jackson, M.J. & Tweeton, D.R. (1994) MIGRATOM, Geophysical Tomography Using Wavefront Migration and Fuzzy Constraints, R.I. 9497, Bureau of Mines, United States Department of Interior, 35 pp. Martin, J., Hardy, M.S.A., Usmani, A.S. & Forde, M.C., (1998) Accuracy of NDE in bridge assessment, Engineering Structures, Vol 20, No. 11, McCann, D.M. & Forde, M.C. (2001) Review of NDT Methods in the Assessment of Concrete and Masonry Structures, NDT&E International, Elsevier Science, Vol 34, 2001, AN

225 Padaratz, I.J., Hardy, M.S.A. & Forde, M.C. (1997), Calibration of radar testing of concrete, Proc. 7th Int. Conf. Structural Faults + Repair-97, Edinburgh, 8-10 July 1997, Engineering Technics Press, Vol. 2, Padaratz, I.J. (1996) A numerical and experimental investigation of radar coupling and propagation through concrete, Ph.D. thesis, University of Edinburgh. Padaratz, I.J. & Forde, M.C. (1995a), A theoretical evaluation of impulse radar wave propagation through concrete, J. Non-destructive Testing & Evaluation, 12, Padaratz, I.J. & Forde, M.C. (1995b), Influence of antenna frequency on impulse radar surveys of concrete structures, Proc. 6th Int. Conf. Structural Faults + Repair-95, London, July 1995, Engineering Technics Press, Vol. 2, Tavlove, A. (1995) Computational Electrodynamics: The Finite Difference Time Domain Method, Artech House (2002) software downloadable free of charge from web site. Yee, K.S. (1966) Numerical Solution of Initial Boundary Problems Involving Maxwell s Equations in Isotropic Media, IEEE Transactions on Antennae and Propagation, Vol 14, AN

226 ADVICE NOTE 3.6 ACOUSTIC EMISSION (AE) Contents Chapter Terminology 1. Introduction 2. Details of Technique 3. Assessing the Condition in Grouted Ducts in Post-Tensioned Concrete: Detection of Wire Fractures in Steel Tendons 4. Testing and Monitoring to Determine the Condition of Reinforced Concrete Bridges 5. Testing and Monitoring to Determine the Condition of Metal Structures 6. Commissioning and Specification 7. Sources of Further Information Appendix I Appendix II Equipment Installation and Verification Case Studies AN.3.6-1

227 TERMINOLOGY Glossary of Terms Acoustic Emission (AE): Term used for transient elastic waves generated by the release of energy within a material or by a process (IS EN :2009, Non- destructive testing-terminology Part 9: Terms used in acoustic emission testing). Terms used to describe the acoustic emission (AE) signal: Acoustic emission signal: The electrical signal obtained from a sensor through the detection of acoustic emission. Attenuation: The observed loss of signal as it travels through a medium. Burst Emission: A description of a signal that has a rapid rise to peak and slower decay. Burst type characteristics are typical of signals collected from fatigue cracks. Event: A single AE source produces a mechanical wave that propagates in all directions in a medium. The AE wave is detected in the form of hits on one or more channels. An event is the group of AE hits received from a single source by two or more channels which can therefore be located. Hit: A hit is the term used to indicate that a given AE channel has detected and processed an acoustic emission transient. Terms relating to the detection of the signal: Feature Data: Data stored in terms of wave parameters such as counts, duration and rise time (see below), which eliminates the necessity of recording actual wave forms, which require a large amount of computer memory and processing. Noise: Signals produced by causes other than acoustic emission, or by acoustic emission sources that are not relevant to the purpose of the test. Terms used to characterise the AE signal: Threshold: The threshold is a preset voltage level, which has to be exceeded before an AE signal is detected and processed. The following terms are made with reference to the threshold (Figure 0.1). Counts: Number of times the signal amplitude exceeds the threshold. Duration: The interval between the first and last time the threshold was exceeded by the signal. Energy (Absolute): The integral of the squared voltage signal divided by the reference resistance (i.e. the resistance of the entire AE measuring system) over the duration of the AE waveform packet. Initiation Frequency: The average frequency of the waveform from the initial threshold crossing to the peak of the AE waveform. Peak Amplitude: Maximum signal amplitude within the duration of the signal. AN.3.6-2

228 Rise Time: The interval between the first threshold crossing and the maximum amplitude of the signal. Terms relating to the equipment: Broadband Sensor: A sensor that has an even response over a large range of frequencies typically (20kHz 1MHz). Couplant: Substance providing an acoustic coupling between the propagation medium and the sensor. DSP: Digital Signal Processing. Hit Based System: A system that records signals on a hit basis and not a cumulative count of signal parameters. Resonant Sensor: A sensor that has a resonant operating frequency and responds across a narrow range of frequencies either side of the resonance. Sensor: Device that converts the physical parameters of the wave into an electrical signal. Figure 0.1 AE Feature Data AN.3.6-3

229 1. INTRODUCTION 1.1 This Advice Note gives guidance on the use of the acoustic emission (AE) technique for: detecting cracking and failure in concrete structures; investigating deterioration in reinforced concrete bridges in particular for providing information on the condition of half joints and hinge joints, and detecting reinforcement corrosion; detecting wire fractures in steel cables in both suspension and cable stayed bridges; also detecting tendon wire breaks in post-tensioned tendons; detecting and locating cracks in welded steel structures. 1.2 The Advice Note also makes reference to the application of AE to investigate bridge bearings, steel corbels and shear studs, and describes the potential use of the technique for giving information on the residual strength of concrete beams. AN.3.6-4

230 2. DETAILS OF THE TECHNIQUE Principle of Method 2.1 An acoustic emission is the release of stored strain energy within a material that generates elastic waves. These waves are generated in every solid body subjected to stress through the occurrence of cracking, corrosion, slip or friction. They propagate through the material and can be detected and recorded by AE sensors mounted on or embedded in the material. Each detected wave is turned into an electrical signal by the sensor and can be subsequently analysed and interpreted. The general principle of AE is illustrated in Figure 2.1. Figure 2.1 Schematic Representation of the AE Principle 2.2 The AE method has two main differences over other non-destructive techniques: firstly, the signal has its origin in the material itself and not in an external source, although it could arise as a result of an external influence such as the loading of a bridge, and secondly it detects movements, not absolute geometry. 2.3 Physically AE waves consist of P-waves (longitudinal), S-waves (shear transverse) and surface waves (Rayleigh waves), which are reflected and diffracted within the material. The P-waves are generally associated with the normal component of stress and a change of volume, unlike S-waves which are associated with the shear component of stress and therefore, with an equivoluminal change. The different velocities of the elastic waves are summarised in the following equations: v R < v S < v P where V R = velocity of surface wave V S = velocity of shear-transverse wave V P = velocity of longitudinal stress wave AN.3.6-5

231 2.4 Typical longitudinal wave velocities for materials commonly used in bridge construction are: 3,500 4,500m/sec for concrete, 5800m/sec for steel and 4,500m/sec for cast iron. These are not necessarily the velocities used for location. For example, in steel structures, the speed of the flexural mode should be used as plate waves are present see Chapter The AE signal is a combination of all these waves and depends on the source, the properties and the geometry of the material as well as on the sensor type. All AE signals contain an element of background noise due to electrical sources (e.g. electromagnetic interference, noise generated by the AE monitoring system itself or by the cables) and due to mechanical sources (e.g. test machines, frictional noise, human activity). Methods employed to increase signal to noise ratio include frequency filtering, increasing the AE system threshold and data analysis. Anti-noise materials such as rubber sheets are impractical in bridge investigations and are more suited to laboratory studies. 2.6 In the case of steel structures, AE from cracks depends on both the speed and size of the fracture. Studies on the characteristics of AE from cracks have shown that it can be separated into primary and secondary emissions. Primary relate to fatigue crack growth and secondary to all other sources. Secondary emissions in all bridge materials include crack face fretting, crack opening and closure, and grinding of debris between the crack faces. 2.7 One of the most important diagnostic effects in AE is the Kaiser effect. This states that, if stresses are applied, removed and then reapplied to a structure, no acoustic emissions occur until the maximum load of the previous stage is reached. Therefore, acoustic emission is irreversible, and this irreversibility or stress memory can be used to determine the magnitude of the previous stress to which the structure has been subjected. 2.8 In some materials, such as concrete, the Kaiser effect is not wholly valid when the level of stress becomes high, acoustic emissions can start below the previous maximum load. This behaviour is called the Felicity effect and the Felicity ratio is defined as: Felicity ratio = load at which emissions start/ previous maximum load 2.9 As a material approaches failure, the Felicity ratio decreases.ae from different materials differs and a brief description of that produced by the main construction materials is given in the sections on application of AE. Test Strategy 2.10 There are several testing strategies that can be employed to evaluate bridge structures. The correct strategy is chosen through consultation with the bridge engineer and the testing contractor in order to identify the test objectives. The expected output from the monitoring and its limitations should be defined. A site visit should normally be undertaken before the work is costed, to assess the structure and the suitability of the proposed strategies, as well as to determine any specific requirements Strategies can be split into two parts: the level of monitoring (number of sensors and information required) and the duration of monitoring. Level of Monitoring 2.12 Global AE monitoring provides a first coarse AE assessment of a structure using AE sensors at a large spacing (the exact spacing is determined from an attenuation survey on site). It can be used to identify the general position of areas of active AE, often using linear location. It is ideal for health checking structures or screening large areas for active damage, provides a fingerprint of condition, and can be used for longer term asset management monitoring. AN.3.6-6

232 2.13 In some cases, active AE locations can be followed up with more sensor-intensive monitoring. If defects are suspected within a structure, it can be beneficial to use a more intensive strategy such as semi- global or local monitoring from the outset to provide accurate source location Semi-global AE monitoring aims to provide a 100% volumetric condition assessment of structures. It is more suited to steel plate structures e.g. box girders, including attachments such as shear studs. AE sensors are placed all over the structure. Individual sources of AE activity are identified and located. Sources are individually assessed, evaluated, and ranked based on source characteristics and on the number of emissions. A follow-up inspection allows monitoring of both structural condition and individual defects allowing the condition of structures to be ranked and the results to be used for longer term asset management/monitoring Local AE monitoring provides detailed information on source location, orientation and on the characteristics of cracking and failure. Testing is often used to confirm the exact location, but it can also be used to monitor known sources for growth rate and changes in characteristics. In the case of steel structures, local monitoring can also be used to assess structural repairs by checking whether the defect s growth has been arrested. This has been used extensively by the Federal Highway Authority (FHWA) in the USA. Duration of Monitoring 2.16 General Considerations. It is crucial that the correct monitoring duration is selected so that active defects may be detected. Key to this is an understanding of the damage mechanisms and an awareness of how and when they are likely to propagate and thus, be identifiable by AE monitoring. A short monitoring duration may miss defects inactive under prevalent conditions whereas extended monitoring may provide no further new information and be carried out at unnecessary cost. Global and semi-global monitoring is more appropriate for steel plate structures, whereas local monitoring is applicable to both steel and the more heavily damped concrete structures Short-term monitoring is used to establish the condition of the structure. The monitoring duration is normally defined by the damage mechanism and is required to be sufficiently long to acquire a representative amount of AE data from active sources to allow satisfactory location and characterisation If the damage is fatigue cracking in a steel structure, then monitoring requires a representative period of vehicular loading e.g. one day on motorway bridges or two or more days for less frequently trafficked bridges. However, if a known large load will be crossing the bridge, then monitoring during that time can be utilised. The latter would be particularly relevant to concrete bridges Permanent monitoring is used when the defect is serious and failure or an increase in damage size would lead to the closure of a structure. Alternatively, permanent monitoring may be used if damage propagation is sporadic with long intervals in between where a short period of monitoring would miss AE activity. Furthermore, permanent monitoring might be appropriate for sub-standard bridges as described in NRA BD79. Examples include detection of wire breaks in suspension cables, cable stays and post-tensioning tendons in post-tensioned concrete bridges Permanent Sensor Mounting (Repeat Monitoring). If a structure requires periodic evaluation to assess degradation it may be cost effective to permanently install AE sensors and signal cables on structures, as this eliminates costly access and labour costs for installation. The monitoring system can then be plugged into the signal cables terminated in a junction box located in a safe and accessible location. AN.3.6-7

233 Accuracy 2.21 The factors affecting the accuracy of AE are defined in the following three subsections which describe detection, location and characterisation of AE signals. Good accuracy requires experienced personnel and an AE system with feature extraction for signal characterisation. Detection Accuracy 2.22 The accuracy of detection of active sources is dependent on: loading (vehicle or environmental) which should be sufficient to propagate damage; the monitoring period which should be of sufficient duration; the detection range of the sensors which should be established using an attenuation survey; the choice and installation of sensors and set up of suitable monitoring equipment (to ensure adequate sensitivity and to increase the signal to noise ratio). Location Accuracy 2.23 The accuracy of source location is dependent on the following factors: The frequency range of the sensors. For steel, one normally uses sensors in the range kHz giving a maximum accuracy of about +/-20mm based on the wavelength of the signals. The higher the frequency, the higher the accuracy. Lower frequency sensors might be chosen for concrete: kHz. Wire break transducers for monitoring post-tensioned structures might be below 50kHz. The monitoring strategy. For global monitoring incorporating a linear array, the source should be in the same plane as the sensors. The level of accuracy and the capabilities of the location strategies should be defined and confirmed with the engineer at an early stage. Planar two- dimensional location on simple plate structures (such as box girders) can have very high location accuracy up to +/-20mm. On concrete structures a less well defined location of cracking or crushing will be achieved. The system set up. This should enable the arrival time to be correctly measured. As in all NDT work, multiple elements, changes in element thickness and the edge of structures cause reflection of signals that can result in inaccurate location of defects or even duplicate sources, i.e. the apparent identification of two sources where only one is present. This problem is particularly important in the case of steel plate structures. Structural geometry. Multiple paths for AE signals will lead to difficulties in interpretation. The effect of geometric dispersion of signals can be limited by an understanding of the structure e.g. the hidden structure of box girders, and by placing AE sensors accordingly. When high accuracy is required for complex structures the number of sensors is generally increased The location accuracy of the monitoring system should be verified before the start of a test and the position of detected sources verified at the end of the test. Both tests should use pencil lead fractures commonly referred to as an HN source (IS EN :2009). AN.3.6-8

234 Characterisation Accuracy 2.25 The accuracy of signal characterisation is dependent on the following: Test operators. Ideally they should be Certified and demonstrate experience of testing similar structures (see para 6.3). Clear procedures (based on verified AE results) for analysing data using defined criteria for accepting, rejecting and classifying data. Accurate source location so that signals from a common source can be analysed Analysis of waveforms may only form a small part of any practical testing and is mostly used during laboratory work. Applications 2.27 AE can be used for a variety of applications including: detection of cracking and failure in concrete structures; detection of deterioration in reinforced concrete bridges in particular for providing information on the condition of half joints and hinge joints, and detecting reinforcement corrosion; detection of giving information on the residual strength of concrete beams; detection of wire fractures in steel cables in both suspension and cable stayed bridges; also detecting tendon wire breaks in post-tensioned concrete bridges tendons; detection and location of cracks in welded steel structures; detection of fatigue cracks in steel box girder bridges; detection of fatigue cracks in shear studs and degradation of concrete surrounding the studs; detection of deterioration in bearings; detection of fatigue cracks in steel corbels Further details are provided in Chapters 3, 4 and 5. Advantages 2.29 Advantages of the AE method include: It can be used remotely to monitor continuously the condition of a structure. It can be applied to structures Globally to screen or health check for defects; this can be done with relatively few sensors. The instrumentation is relatively quick and easy to install and can be linked to a modem to provide remote data collection. When used for monitoring structures under normal loading, it requires only limited access to the structure. It can be retrofitted to existing structures, adapted to existing topologies and is not invasive. AN.3.6-9

235 It can locate the sources generating elastic waves and can be used to determine their approximate crack locations before they are visible on the surface. It is the only technique that can detect and locate active flaws such as growing cracks, irrespective of their orientation within the structure; this can be done in real time. It can uniquely identify active defects many metres away so prior knowledge of their location is not required. It can be used to detect and locate hidden defects inside structures (e.g. internal welds), in areas obscured from view or in areas that are difficult to inspect, as well as under paint or dirt. Semi-global and local monitoring can accurately locate defects, in some situations to within +/- 20mm. It can be used in conjunction with parametric data (strain, displacement, temperature, etc) to investigate their relationship with the AE. Once the damage mechanism is understood, AE can be used for either: short-term periodic monitoring (inspection); long-term continuous condition monitoring. It can be used at a strategic level to rank groups of structures or areas of structures based on condition. It can be used at an operational level to assist with the inspection of specific structures. Limitations 2.30 The limitations of the AE method include: Equipment It requires expensive specialist instrumentation which is hit based and has individual data processing channels. Successful AE monitoring requires specialised, trained, certified and experienced personnel to set up the instrumentation, acquire the data on- site and to analyse the data using robust procedures. It does not detect defects that are either no longer propagating or not growing during the monitoring period, although it does detect fretting of cracks. The absence of AE does not mean that a structure is in good condition, for example, there may be dormant cracks which could become active if, for example, the structure was subjected to increased loading. Therefore, a structure should be monitored under representative peak loading. It can only be used to obtain an approximate measure of the active length of defects and other techniques such as ultrasonics (where applicable) should be used It may not be applicable where the critical crack size is very small and failure is sudden, e.g., in cast iron structures The emissions depend on the load configuration and on the material under test and are affected by attenuation and background acoustic noise A typical AE system comprises a high speed DSP, AE processing boards with individual processing channels for each sensor (i.e. a non-multiplexing system) and the ability to programme the settings for signal thresholds and frequency range to enable the AE signal to be filtered. It should also have AN

236 software for source location in both one and two dimensions, feature extraction capability to allow characterisation of the signals and stable software for long-term monitoring. Two typical systems are shown in Figure 2.2). Figure 2.2 Typical AE Monitoring Systems 2.32 It should be calibrated annually (against the NSAI standard or equivalent) and checked before use in the field An AE signal is recorded when the signal reaching the sensor exceeds a threshold value, set by the operator in order to avoid recording unwanted noise (e.g. 30 to 45 db is generally satisfactory for concrete) AE sensors are piezoelectric crystals that convert movement (a variation of pressure) into an electrical voltage (Figure 2.3). The sensors must all have an identical response and they should be calibrated annually. They are normally held in place using metallic clamps (see Appendix I) and connected to the AE system using coaxial cables with shielding to prevent electro-magnetic interference There are two types of AE sensor: resonant and broadband. Resonant sensors are more sensitive to sources at particular frequencies thus sensor spacing can be maximised and they help minimise background noise by filtering out unwanted frequencies. Sensors with a low resonant frequency are more susceptible to noise but high frequency signals are readily attenuated which limits their use. A resonant frequency of kHz is normal for concrete applications; whereas 100 and 200kHz is used for metallic structures. Higher frequency sensors can be used in high noise environments but only for local monitoring due to the higher attenuation at these frequencies Broadband sensors are more suited to source characterisation and fundamental studies in the laboratory due to the bandpass filtering in resonant sensors removing parts of the original signal. Broadband sensors, however, have a reduced sensitivity, which requires closer sensor spacing, increasing the cost. AN

237 Figure 2.3 Typical AE Sensors Figure 2.4 Proprietary System and Sensor AN

238 2.37 The following additional facilities may be required when using an AE system on a bridge site: Environmental protection for the monitoring system. A mobile elevated working platform (MEWP) or scaffolding for access to the structure v AC power. A phone line or ISDN line for computer communication (if remote monitoring). Traffic management. General site lighting. Security. Welfare facilities for personnel on site Although there are a number of companies producing AE equipment only a very small number actually provide a testing service and bridge testing should not be undertaken without the relevant training, experience and records 1. Pre-Test Procedure 2.39 The application of AE monitoring should include the following steps: Aim of the Test The aim of the test should be defined and each of the following steps should then be undertaken. The procedure adopted may need to be modified during the test to take account of the results obtained. Preliminary Survey A preliminary survey should be carried out, comprising a review of the structure s records including as-built drawings, photographs, records and maps of existing damage and previous monitoring data, if available. Selection of AE System It needs to be a system that enables waveform feature recognition and requires separate data channels (rather than multiplexing). Systems must be able to measure individual signal amplitude (dbae), risetime, duration, absolute energy and counts, as an absolute minimum. They should also have a sufficient sampling rate for feature extraction and waveform analysis. Selection of Sensor It depends on the: material, geometry and size of the structure under test; working environment; type of data analysis required; sensitivity and frequency range required; signal attenuation in the material under test. The sensor selected should have an operating frequency, bandwidth and sensitivity that takes account of these requirements. When a large sensor spacing is needed (i.e. during global monitoring) low frequency resonant sensors are preferred, as they are less influenced by attenuation and are more 1 A list of companies undertaking AE can be found on AN

239 sensitive. If smaller spacings are appropriate or a frequency analysis is required, the broadband type should be used. Sensor Location The position and separation of the sensors depends on the structure itself and the aim of the monitoring. Global monitoring of the whole structure generally implies a large spacing, whilst local monitoring involves a closer spacing concentrated on a smaller area. The available number of sensors is also an issue as in some cases the transducers are expensive and therefore, can affect the cost-effectiveness of an investigation. Sensor Mounting and Coupling The method used to mount the AE sensors on the structure affects the quality and quantity of data that will be recorded and therefore, it has to be carefully considered. The choice of the right couplant is fundamental. The use of clamps to hold the sensors in place is also advisable. Cables should also be held in position to avoid interference from electromagnetic signals induced in the cables by movement due to wind. Instrumentation Settings The settings of the AE instrumentation depend on the type of system and the software that is used. If the data are stored locally, the first check should be on the amount of disk space available. Waveform data in particular generate large size data files. If the data are transferred to a data centre, the necessary connections have to be provided. Although the settings are decided at this stage, results and/or considerations that arise during the calibration and noiseattenuation studies, might give reason to modify these initial settings. A record should be kept of the settings used during the test. Mounted Sensor Sensitivity Check Once the sensors are chosen, located and mounted they need to be verified. A standard procedure for checking sensitivity is described in IS EN :2009. It requires a lead pencil to be broken 50mm from each sensor and provides confidence in the sensitivity of the transducers. The check should be done at the beginning, at the end and if possible in the middle of the test to assure correct performance throughout the whole test. Noise and Attenuation Study When working in the field, a noise and attenuation study can provide useful preliminary information about the condition of the structure under investigation. Further details are provided in Appendix I. Parametric Data And Environmental Conditions During AE monitoring, it may be useful to record additional information, such as displacement and/or strain together with the environmental conditions under which the measurements were made such as wind, temperature and rain. These can significantly affect the measurements and should be recorded - as well as considered when setting the threshold value of the equipment. For example, wind could cause movement in data cables causing some emission. Testing Procedure 2.40 A detailed description of the procedure for monitoring a bridge is given in Appendix I. AN

240 3. ASSESSING THE CONDITIONS IN GROUTED DUCTS IN POST-TENSIONED CONCRETE: DETECTION OF WIRE FRACTURES IN STEEL TENDONS Principle of Application 3.1 When a highly stressed steel wire fractures there is a release of energy, which is transmitted along the wire or into any structure to which it is attached. These acoustic events can be detected by AE sensors attached to the external surface of the structure. 3.2 This approach was initially used on buildings for detecting wire breaks in post-tensioned floor slabs containing unbonded tendons. The operation of the system on a bridge with grouted tendons was believed to be more difficult for two reasons. Grouted tendons would probably release less energy on fracture than unbonded tendons, and non-break events and background noise were likely to be more dominant. However, trials at TRL have demonstrated that the method can be applied to steel tendons in grouted post- tensioned ducts. Accuracy 3.3 The following table gives an indication of the success of the technique in detecting wire fractures in trials on bonded post-tensioned tendons. It also summarises the results of trials on a hanger cable and the main cable of a suspension bridge. The fractures were created in various ways as described below. For the post-tensioned beams, external wire breaks were created by attaching a rig containing a stressed strand to the concrete surface. To create a wire fracture, a notch was cut in one of the wires and a load applied to the strand and increased until the wire fractured. The facsimile wire break events were created by a Schmidt hammer or spring impactor applied to the surface of the concrete. The wires in the hanger cable fractured whilst under cyclic loading and those in the suspension bridge cable were cut mechanically. Test Nature of trials Outcome Post-tensioned beam External wire breaks 25 breaks, five blind, all detected, 22 located within 0.2m and 3 within 0.5m of the fracture. Post-tensioned bridge External wire breaks and facsimiles 44 wire breaks and facsimiles, 18 blind; 43 detected 18 within 0.2m and rest within 0.5m, 41 events correctly classified. Bridge hanger cable Fatigue test fractures Nine wire fractures were found but eleven were detected, i.e. two were erroneous. Bronx Whitestone suspension bridge main cable Wires cut mechanically Six wires cut blind, five detected within 0.22m and one within 0.7m. Table 3.1 Summary of Results from Trials AN

241 Applications 3.4 In addition to detecting wire fractures in steel tendons in grouted ducts in post-tensioned concrete bridges, it can also be used to detect wire fractures in: hanger cables in suspension bridges; suspension bridge cables. Advantages 3.5 AE is the only NDT technique that can reliably detect wire fractures in post-tensioned tendons in grouted ducts. 3.6 AE does not require holes to be drilled into the concrete for inspection. Such holes can allow oxygen into the void which increases the risk of corrosion. Limitations 3.7 Care is required to ensure that the system is set up to distinguish between wire fractures and other sources of AE that have similar characteristics. 3.8 It is only possible to detect wire fractures that occur during the period that the structure is being monitored. It is not possible to detect fractures that have already occurred. Equipment and Resources 3.9 The detection of wire fractures can be achieved using standard AE systems which are configured for this application or bespoke systems which have been specially designed for this purpose. Commercial bespoke systems employ hardware and software filters to reject AE signals of no interest such as those due to vehicles going over expansion joints and discontinuities in the road surface, and small objects such as stone chippings striking the concrete. Events that pass these filters are recorded as wire fractures. Evaluation Trials 3.10 Unless a monitoring system has previously been proved to work in a comparable location, evaluation trials are desirable to: confirm that wire breaks can be detected in that particular application; demonstrate they can be distinguished from the non-break events expected on the site; show that they can be captured reliably in the presence of the ambient noise. Interpretation 3.11 If a bespoke system is used or the AE system is properly configured, interpretation is straightforward as the system reports the occurrence of wire breaks. AN

242 4. TESTING AND MONITORING TO DETERMINE THE CONDITION OF REINFOCED CONCRETE BRIDGES Principle of Application 4.1 The method can be used to obtain information on the condition of reinforced concrete bridges, in particular the condition of half joints and hinge joints. The basic principles are described in Chapter The Kaiser effect has been shown to exist on concrete beams subjected to tensile cracking as long as the crack width is less than mm. The effect breaks down when shear cracks appear and start to play a primary role. 4.3 The accuracy, advantages and limitations of the technique when applied to concrete are described in Chapter 2. Applications Application Sources of AE Outcome of Tests Half-joints Hinge-joints Load carrying capacity of beams Reinforcement corrosion Corrosion, delamination, spalling, fatigue cracks in reinforcement bars Corrosion, delamination, spalling, fatigue cracks in reinforcement bars Application of load by load test or traffic Corrosion, delamination, microcracking of concrete due to corroded material expansion A field trial investigated eight half joint nibs and ranked the joints in terms of detected energy. The most active joint was visually the most degraded and had a high chloride contamination. Fatigue cracks were detected in laboratory investigations of reinforcement bars encased in concrete. Field trials have been used to rank bridges in terms of detected energy and shown degradation of concrete around reinforcement bars. Possibility of providing information on residual load capacity. Both laboratory and field investigations have shown that corrosion is detectable with AE. Table 4.1 Example Applications of AE Investigations in Concrete Structures AN

243 Equipment and Resources 4.4 The equipment used is described in Chapter 2, although the working frequency range is normally limited to kHz as higher frequencies are rapidly attenuated. Sensor Spacing. Generally, a maximum separation of 2m is reasonable when using resonant low frequency sensors (60kHz). This distance has to be reduced by approximately half if using broadband sensors. The presence of pre-existing cracks or damage can affect the signal and its propagation and, therefore, has to be taken into account when deciding the sensor spacing. Mounted Sensor Sensitivity Check. The Schmidt Hammer or a centre punch are quick and easy ways to verify that all the transducers are functioning, whilst a pencil lead break near to each sensor provides confidence on the sensitivity of the transducers, as described in Chapter 2. Interpretation 4.5 Interpretation of the AE signals should include a simple source location to identify the most active areas. An analysis of the AE hit rate, rise time and duration can provide further information, for example, it has been observed that: microcracks generate a large number of events of small amplitude whilst macrocracks generate fewer events but of larger amplitude; when the cracks open up, most of the energy is released and subsequently, many small amplitude events are created. 4.6 The results can be used to determine whether the structure is in a stable condition or deteriorating and how it compares with other structures of a similar type. 4.7 Tests on concrete specimens under load have shown that when load is first applied, there is an initial burst of AE followed by a decrease in activity as the load is increased further and then an increase in activity just before the failure. This phenomenon was named the Silence effect and its duration the Silence time. An explanation for this has been proposed based on the Griffith theory of fracture. According to this theory, numerous small cracks are present during the initial stage of the fracture process giving rise to several AE sources. Afterwards, when the damage localises, the cracks join up resulting in fewer AE sources and consequently less AE activity. 4.8 A simple criterion that has been introduced by the Japanese Society for Non-destructive Investigation to quantitatively assess concrete structures is based on the following two parameters (Figure 4.1): Load ratio = load at the onset of AE activity in the subsequent loading/the previous maximum load. Calm ratio = the number of cumulative AE hits during the unloading process/total AE activity during the last loading cycle up to the maximum load. 4.9 The parameters can be used to classify the degree of damage as shown by the dashed lines in Figure 4.1. The position of the lines needs to be determined in advance, based on preliminary tests, in order for this approach to be used for practical applications. AN

244 Figure 4.1 Classification of damage according to NDIS from [Ohtsu et al., 2002] 4.10 The generation of AE activity during unloading is an indication of structural instability. For a structure in good condition, no acoustic emission is generally recorded during the unloading phase. Based on this concept a new parameter, named Relaxation ratio, was developed and a new type of analysis has been proposed, as illustrated in Case Study Some more sophisticated forms of analysis are described in the case studies. AN

245 5. TESTING AND MONITORING TO DETERMINE THE CONDITION OF METAL STRUCTURES Method of Application 5.1 The application of AE to metal bridges results from decades of use on metal structures, historically from the petrochemical industry. The first published record of a bridge test was in the UK on an army Bailey bridge in Since then the technique has been widely used in the USA by the Federal Highway Administration and has also been used on trunk road bridges in the UK. 5.2 The primary application of AE in metal structures is welded steel such as steel box girders, orthotropic bridge decks and welded I-beam girders; however, other bridge elements can be monitored such as steel corbels, shear studs and bridge bearings. 5.3 When applied correctly, AE monitoring detects, locates and quantifies active defects such as steel fatigue, and also detects changes in condition over time. As well as detecting fatigue cracks, it is possible to detect their secondary effects such as plastic deformation and crack face fretting (rubbing). See Figure 5.1. It should be noted, however, that, in time, existing crack interfaces can become smooth so that emissions from fretting can disappear. Lack of emissions can therefore disguise the existence of dormant cracks. 5.4 The AE technique has been used successfully in the laboratory to locate fatigue cracks in girder welds, models of riveted railway bridges, steel reinforcing bars and shear studs. Field applications have demonstrated the use of AE to detect fatigue like AE sources in box girders, shear studs and welded I-beams. AE sources on welded I-beams and roller bearings have been positively identified as active cracks. 5.5 The advantages and limitations of the use of AE monitoring on metal bridges and the equipment and resources required are described in Chapter 2. Accuracy 5.6 The detection of fatigue cracks in metal structures relies heavily on the quality, specification and installation of the AE equipment. It has been shown in laboratory studies that it is possible to detect and locate fatigue cracks with a visible length of less than 2mm. AN

246 Application Application Sources of AE Outcome of Tests Box Girder Bridges Shear Studs Bearings Steel Corbels Fatigue cracks in parent and weld metal Fatigue cracks in shear studs and degradation of concrete surrounding stud Fatigue cracks, damage to lubrication and sliding surfaces Fatigue cracks in parent and weld material Both laboratory and field investigations have proved successful and sources indicative of fatigue cracks have been detected and located. Laboratory investigations were used successfully to detect and locate a fatigue crack in shear studs (verified at 25% crosssectional area). AE has been successfully used to assess the condition and operation of a bridge bearing. AE results have been used to justify keeping a bridge structure in service. AE has been successfully used to show the integrity of steel corbels. Table 5.1 Example Applications of AE Investigations in Metal Structures 5.7 Examples of the above applications are presented in Appendix II. Figure 5.1 AE Signal Differentiation in the Laboratory between Crack Propagation and Crack Fretting AN

247 6. COMMISSIONING AND SPECIFICATION Introduction 6.1 The following are specific requirements for the use of AE and should be used in conjunction with those given in Advice Note 1: General Guidance of this series of Advice Notes. 6.2 Information Supplied to Tenderers 6.3 The tender documents submitted to testing organisations should state the objectives of the test and the level and duration of monitoring that is required. If the former, details should be given of the test. If the latter, an indication should be given of the period of monitoring. In addition, the following information should be included: Any available information about the actual known condition of the structure. Any available information about previous AE testing and/or monitoring. Information Required of Tenderers 6.4 The tenderers should be asked to provide the following additional information: The specification for the AE system that is proposed, i.e. it s make and model, type of acquisition system, e.g. hit based acquisition, feature extraction capability and type of source location provided. The number, type and specification of the AE sensors that will be used. An indication of where the sensors will be located, how they will be mounted, their separation and the procedure for checking their sensitivity when installed. A general indication of the type of analysis that will be undertaken. Details of relevant experience and background in testing. The accuracy and reliability that are anticipated. References of previous work. Details of the qualifications, training and experience of the testing engineers (i.e. to ASNT TC-1A Level 2 AE or forthcoming BiNDT PCN Level 2 AE Certification). Procedures for reviewing the test results. It is suggested that this is done by more experienced personnel with a qualification such as the ASNT Level 3 (or equivalent). Details of their quality system used (IS EN ISO 9001:2009). How to Use the Results Additional Requirements 6.5 The complementary use of other NDT techniques should be considered to confirm the findings obtained from the AE investigation. 6.6 The raw data should be made available if requested. AN

248 Model Output Report 6.7 A final report on the condition of the structure which should include: The make, model and frequency range of the proposed sensors along with a calibration certificate for a typical sensor. The method used to hold the sensor to the structure. The method of coupling the sensor. If the couplant is also the method of attachment limitations on loading should also be included to allow confidence in preventing sensors falling into live traffic The method for verifying sensor sensitivity, location accuracy and attenuation of the structure. Make and model of AE data capturing system. 6.8 It is difficult to include AE set-up information for an acoustic emission bridge investigation for comparison with another investigation completed by a different company. Commercial hardware systems do not all work in the same manner and have different system set-ups for recording the necessary information. The following, however, is universal and should be included: The threshold used for recording signals and whether AE waveforms were also recorded and the corresponding threshold. The type and accuracy of source location and whether location accuracy is based on an experimental investigation on the structure or previous investigations, including references. The background noise level. If background noise were unacceptable, details of the methods used to control it, e.g. guard sensors, using location analysis to exclude known sources of noise from the analysis, or other established methods. AN

249 7. SOURCES OF FURTHER INFORMATION Standards National Roads Authority, BD 79, Management of Substandard Road Structures Design Manual for Roads and Bridges. IS EN Non-destructive testing. Acoustic emission. General Principles. IS EN Non-destructive testing. Acoustic emission. Equipment characterisation. Equipment Description. IS EN Non-destructive testing. Acoustic emission. Equipment characterisation. Verification of operating characteristics. IS EN Non-destructive testing. Terminology. Terms used in acoustic emission testing. ASTM E , 2005, Standard Guide for Acoustic Emission System Performance Verification. ASTM E650-97, 2002, Standard Guide for Mounting Piezoelectric Acoustic Emission Sensors. ASTM E976-84, 2001, Standard Guide for Determining the Reproducibility of AE Sensor Response. ASTM E , 1992, Standard Method for Primary Calibration of AE Sensors. ISO 12713, 1998, Non-destructive testing- acoustic emission- primary calibration of transducers. ISO 12716, 2001, Non-destructive testing- acoustic emission- vocabulary. Detection of Wire Fractures Cullington, DW, MacNeil, D, Paulson, P & Elliott, J (1999) Continuous acoustic monitoring of grouted post- tensioned concrete bridges, Proc. 8th Int. Conf. Structural Faults & Repair-99, Commonwealth Institute, London, July 1999, Engineering Technics Press, ISBN Hill, ME, Bradbury, T & Cullington, DW (2002). Acoustic monitoring of concrete bridges TRL beam trials. TRL Unpublished Report PR/IS/05/02. Paulson, P O & de Wit, M (2003) Use of acoustic monitoring to manage concrete structures, Proc. 10th Int. Conf. Structural Faults & Repair-2003, Commonwealth Institute, London, 1-3 July 2003, Engineering Technics Press, CD-Rom, ISBN AN

250 Concrete Structures Colombo, S., and Forde, M.C. (2001). AE experiments on concrete beams: general overview and research in progress on bridges, Proc. of the Int. Conf. Structural Faults+Repair-2001, M.C. Forde Ed., CD-Rom, July, London. Colombo, S, Forde, M.C., Main, I.G. & Halliday, J. (2003). AE monitoring of concrete bridge beams in-situ, The Structural Engineer, Vol 81, No 23/24, Colombo, S., Main, I.G., Forde, M.C. (2003). Assessing Damage of Reinforced Concrete Beam using b-value. Analysis of Acoustic Emission Signals, Journal of Materials in Civil Engineering, ASCE, May/June 2003, Vol.15, No.3, Colombo, S., Main, I.G., Forde, M. C. and Halliday, J. (2003). Estimating damage and load carrying capacity of concrete bridge beams by AE. Proc. of the Int. Conf. Structural Faults+Repair-2003, M.C. Forde Ed., CD-Rom, July, London. Colombo, S., Main, I.G., Forde, M. C. and Halliday, J. (2002). AE on bridges: experiments on concrete beams. In Mazal P., editor, Proc of the EWGAE th European Conf. on Acoustic Emission Testing, Volume I, , Prague, Czech Republic. Czech Society for Non-destructive Testing. Li, Z., Xi, Y.,(1995) Application of Acoustic Emission Technique to Detection of Concrete Cracking and Rebar Corrosion, NDT-CE: Int. Symposium Non-Destructive Testing in Civil Engineering, 28 Sept.1995, Berlin, Germany, Lyons, R., Ing, M., and Austin, S. A. (2005). Influence of diurnal and seasonal temperature variations on the detection of corrosion in reinforced concrete by acoustic emission. Corrosion Science, 47, No. 2. Feb Nesvijski, E.G. (1997). Failure Forecast and the Acoustic Emission Silence Effect in Concrete. ASNT s Spring Conference. Houston, Texas Ohtsu, M. (1987). Determination of crack orientation by acoustic emission. Materials Evaluation, 45(9): Ohtsu, M., Uchida, M., Okamoto, T. and Yuyama, S. (2002). Damage assessment of reinforced concrete beams qualified by acoustic emission. ACI Structural Journal, 99(04): Pullin, R., Holford, K.M., Lark, R.J. and Beck P., (2003), Acoustic emission assessment of concrete hinge joints. Damage Assessment of Structures. Key Engineering Materials 4, , Pullin, R., Holford, K.M. and Lark, R.J., (2004), An investigation of the use of acoustic emission to monitor hinge joints, Cardiff School of Engineering Report No. 3060, Feb Yuyama, S., Imanaka, T., and Ohtsu, M., (1988), Qualitative evaluation of microfracture due to disbonding by waveform analysis of acoustic emission, The Journal of the Acoustical Society of America, 83 (3) Yuyama, S., Okamoto, T. and Nagataki, S., (1992), Acoustic emission evaluation of structural integrity in repaired concrete beams, Materials Evaluation, 52(1), AN

251 Yuyama, S., Okamoto, T., Shigeishi, M., Ohtsu, M., Kishi, T. (1998). A proposed standard for evaluating structural integrity of reinforced concrete beams by acoustic emission. Acoustic Emission: Standards and Technology Update, ASTM STP American Society for Testing and Materials. S.J. Vahavilos Ed., Metal Structures Carter, D., (1999). Acoustic Emission Technique for the Structural Integrity Monitoring of Steel Bridges, Post Graduate Research Thesis, Cardiff University. Kaiser Von J. (1953). Knowledge and Research on noise measurements during the tensile stressing of metals. Archiv fur das Eisenhuttenwesen. 24(1335): Jan-Feb. (in German). Pullin, R., (2001). Structural Integrity Monitoring of Steel Bridges Using Acoustic Emission Techniques, Post Graduate Research Thesis, Cardiff University. Pollock, A.A. (1989). Acoustic Emission Inspection. Metals Handbooks. Vol. 17 ASM International. 9th Edition Pullin, R. Carter, D.C., Holford, K.M. and Davies, A.W., (1999). Bridge Integrity Assessment by Acoustic Emission Local Monitoring. 2nd International Conference on Identification in Engineering Systems, Watson, J.R., Yuyama, S., Pullin R. and Ing M., (2005). Acoustic Emission Monitoring Applications for Civil Structures. Surrey University s International Bridge Management Conference. Watson, J.R. et al, (2001). Remote Detection of Damage in Bridges, Structural Faults and Repair Conference Watson, J. R., (2000)a. BOXMAP - Non-Invasive Detection of Cracks in Steel Box Girders, University of Surreys 4th International Bridge Management Conference, 17th-19th April 2000, Guilford, Surrey. Watson, J.R., (2000)b. Condition Monitoring of Steel Bridges Using Acoustic Emission, Post Graduate Research Thesis, Cardiff University. Other Publications Carter, D. and Holford, K.M., (1996), I.M.A.G.IN.E.: Letting bridges do the talking, Insight,. 38 (11), Carter, D.C. and Holford, K.M., (1998). Strategic Considerations for the AE Monitoring of Bridgesa Discussion and Case Study, INSIGHT, Vol. 40 No. 2, 1-5. Henning, D. (1988). Josef Kaiser: His Achievements in Acoustic Emission Research, Materials Evaluation, 46: , Feb. AN

252 Holford, K.M., Carter, D.C., Pullin, R. and Davies, A.W., (1999). Bridge Integrity Assessment by Acoustic Emission Global Monitoring 2nd International Conference on Identification in Engineering Systems, Mori,Y., Obata, Y. (1998). Characteristics of Acoustic Emission source in a fatigue crack. Nondestructive Testing Communications. 4: Physical Acoustics Corporation, (1995). Acoustic Emission For Bridge Inspection Report No. FHWA- RD- 94- prepared for FHWA and U.S. Department of Transportation, June Pollock, A.A. and Smith, B., (1972). Stress Wave Emission Monitoring of a Military Bridge, Nondestructive Testing, Vol 30, No 12, Prine, D.W., Hopwood, T., (1985), Improved Structural Monitoring with Acoustic Emission Pattern Recognition, Proceedings of the Fourteenth Symposium on Nondestructive Evaluation. San Antonio, TX. Sammonds, P.R., Meredith, P.G., Murrel, S.A.F., Main, I.G., (1994) Modelling the damage evolution in rock containing pore fluid by acoustic emission, Eurock 94, Balkerna, Rotterdam. Scott, I. G. (1991). Basic Acoustic Emission - Non Destructive testing monographs and tracts. Gordon and Breach Science Publishers Shearer, P.M.(1999), Introduction to Seismology. Cambridge University Press, Shigeishi, M., Colombo, S., Broughton, K.J., Rutledge, H., Batchelor, A.J. and Forde, M.C. (2000), Acoustic Emission to assess and monitor the integrity of bridges, Construction and Building Materials, Vol.15, Shigeishi, M., Masaki, Y., Jo, H., Fujimoto, S., Makizumi, T., and Matsushita, H. (1999). Acoustic Emission on a 60 years old bridge beam under bending test. In Proc. of the Int. Conf. on the current and future trends in bridge design construction and aesthetics, , Singapore. Shinomiya, M., Nakanishi, Y., Morishima, H., Shiotani, T. (2002). Damage diagnosis technique for brick structures using Acoustic Emission. Proc. of the EWGAE th European Conf. on Acoustic Emission Testing. Ed. P. Mazal. Czech Society for Non- destructive Testing Sept. Prague, Czech Republic., Vallen, H., (2002), AE Testing Fundamentals, Equipment, Applications, NDT.net - September 2002, 7 (09). AN

253 APPENDIX I: TESTING PROCEDURE Installation AI.1 AI.2 AI.3 AI.4 AI.5 AI.6 AI.7 Sensors are normally mounted to the external surface of the structure with the area to be evaluated inside the sensor array. The installer needs to be in close proximity to the surface of the structure at sensor locations, thus, sensors are normally installed from a mobile elevated work platform (MEWP) or scaffolding. Installation may require temporary road closure, traffic management or night time working. The surface of the structure should be of sound material (no loose rust, paint or concrete) and smooth with no irregularities which would prevent complete sensor face contact to the surface resulting in loss of transmission of the signals. In steel structures this can be achieved by light sanding or scraping the paint coating on the bridge structure over an area of 50mm 2 per sensor. It is not necessary to remove all the paint. In concrete structures the surface can be smoothed using a sanding block. A thin layer of couplant (~5ml) is applied to the face of the sensor to allow transmission of signal between the surface and the sensor. Non-viscous, temperature stable multi-purpose greases have been used successfully for steel bridge testing whilst silicon sealant has been used successfully on concrete structures. These are recommended as they have excellent signal transmission properties and are inert. A magnetic clamp should be used to hold the sensor in place on steel structures. This should incorporate a spring which ensures a constant pressure is applied between the structure and the sensor and prevents the internal piezoelectric crystal from being permanently deformed. The sensor instrumentation is shown in Figure AI.1. On concrete structures, an aluminium clamp can be used (Figure AI.2) to secure the sensor to the structure. The rubber mount protects the piezoelectric crystal from being permanently deformed when the clamp is screwed to the structure. Figure AI.1 Sensor Mounting on Steel Structures AN.3.6-I/1

254 AI.8 If the instrumented structure is over live traffic, the clamps and cables should be tied to the structure. The signal cables must be attached flush to the structure to prevent either snagging by passing vehicles or cables knocking the structure during monitoring which causes emissions and to ensure that there is no strain in the cable. This is normally achieved by tying cables to structural details or parapets using cable ties or by using cable tie mounts in concrete structures (FigureAI.2). Figure AI.2 Attachment to Concrete Structures AI.9 AI.10 AI.11 Signal cables should be run along the structure to a convenient position, normally at the end of a span and taken to the monitoring system. Cables can be up to 1000m in length without any appreciable signal attenuation but financially and practically, it is best to try not to exceed 100m. Cable length for all sensors should be the same to avoid differences in signal loss from the sensor to the system. For short-term monitoring, the monitoring system requires environmental protection and should be located in a temporary work site which could be a trailer, van or temporary structure such as a port-a- cabin. In some situations the bridge may provide a suitable environment e.g. inside large box girders, abutments or piers. It is recommended by the equipment suppliers that the environment in which the equipment is housed must be maintained in the temperature range 10-25ºC. Continuous stable power, normally AC, is required for the monitoring system either at 110v or 240v if allowable. Long-term unattended systems require secure housing to prevent vandalism and theft. A phone line dial up connection or ISDN (or equivalent) should be available if remote access is required. System Verification AI.12 AI.13 System verification is essential to ensure that the system is correctly installed. The verification should adhere to the following procedure. The process applies to both concrete and steel structures. A mounted sensor sensitivity check should be completed before any measurements are made. This includes an attenuation survey, wave speed determination and a source location check. The sensitivity is verified using an artificial source. The current approach is to use the pencil lead fracture technique (PLF) (ASTM E976). A PLF 50mm from a sensor that is well mounted, should record a signal between 95 and 97dB. AN.3.6-I/2

255 AI.14 An attenuation survey should be performed to determine the optimum sensor spacing. This can be done by attaching a number of sensors at known spacings as shown in Figure AI.3. Signals from a PLF outside the sensors array should be recorded at all the sensors and a plot of signal amplitude at each of the sensors against distance from the source provides the attenuation of the structure. An example of an attenuation plot in a steel bridge is shown in Figure AI.4. It can be used to ensure that signals from defects can be detected. For example, fatigue cracks in steel have a typical source amplitude of 70dB, the attenuation plot can, therefore, be used to set a spacing that ensures that a signal of 70dB will not fall below the threshold before arriving at a sensor. Special care should be considered at sections containing irregularities such as bolted joints or changes in material thickness as these add to the attenuation of signals. Individual attenuation studies should be completed around these sections and sensor spacing reduced accordingly. Figure AI.3 Instrumentation for Determining Attenuation of Structure Figure AI.4 Example of Signal Attenuation in a Steel Bridge in the Khz Frequency Range AI.15 For accurate source location, a measurement of the wave velocity must be determined. The wave velocity can be measured by fracturing a pencil lead outside the sensor array and recording its time of arrival at two sensors as shown in Figure AI.5. The wave speed can be determined from the arrival times and used for all location analyses. AN.3.6-I/3

256 Figure AI.5: Determination of Wave Speed AI.16 The location accuracy of the installed system should be determined. This can be done by making a PLF at known positions on the bridge and comparing its location with that determined by the AE system from the arrival times of the signals from the PFL. This should be done using an increased threshold to compensate for the PLF amplitude being greater than the typical amplitude of fatigue signals. The location of sources identified during monitoring can be checked using the PLF technique after the completion of the test. If possible, the location of significant sources should be measured relative to a datum point on the structure, and a permanent mark left to allow easy identification. Noise Checks AI.17 AI.18 AE noise sources from bridges include deteriorated bolted joints, bearings and the effects of traffic loading. Noise in structures that are causing problems can be removed using either a location technique or guard sensors. Signals from areas of noise can be located and not recorded during the monitoring process. Guard sensors can be positioned around a noise source and if any of these sensors are the first hit sensor, any signals they record can be discarded during recording or during analysis. This is achieved by proper system set up and from information gained during the initial structural calibration carried out at the start of each test. An analysis of ambient noise sources in any bridge should be completed prior to monitoring. A discrete period of monitoring should be completed whilst there is no loading on the structure to establish that there are no extraneous noises present. This can be achieved by recording data whilst visually observing no traffic loading on the structure. This process can be described as the evaluation of background noise. The threshold of the system should be set above the threshold of the background noise. Parametric Data AI.19 The load applied to the bridge should be monitored and compared with the AE output. A suitable approach would be to use a linear voltage displacement transducer or strain gauge(s) at critical locations to monitor deflection or strain, respectively and hence, loading, and compare that with recorded AE. It is beneficial to include a period of visual traffic monitoring to allow comparison of vehicle size with recorded AE. Test Records AI.20 An AE test log should be completed during the test, ideally in test specific worksheets. The log should be retained and archived as a key document for each test. The log should contain relevant information on engineers and their relevant qualifications and experience, test files and where they are stored and a time log of the test indicating points of interest or problems that occurred during the investigation. The ambient temperature and weather conditions should also be recorded. AN.3.6-I/4

257 Data Analysis and Interpretation AI.21 Once all the data (AE, environmental and parametric measurements, e.g. strain, displacement) have been collected, they need to be processed and interpreted. The method used to process the data will depend on the aim of the investigation and type of recorded data. Some examples are given in Appendix II. The procedures used must be derived from representative and well conducted laboratory work or field experience where damage has been verified and correlated by other means than AE. AI.22 AI.23 AI.24 AI.25 AI.26 AI.27 AI.28 AI.29 AI.30 Different types of defects generate different types of AE signals with varying frequency ranges and amplitudes. These differences can sometimes be related to the degree of damage of the structure. For example, signals from fatigue cracks should have a burst type appearance. AE signals can be analysed in an analogue or digital manner. An analogue analysis consists of a parametric study of several AE attributes (i.e. amplitude, duration, AE counts, rise time see Figure 0.1), which can be correlated with the fracture process and/or with the properties and behaviour of a structure. A digital analysis comprises a waveform analysis which examines the characteristics of the whole AE wavetrain such as frequency content and wave shape. Recording waveforms for discrete periods during an investigation can be used to characterise any located source, however this is not practical for the majority of field work. AE waveforms and characteristics can be compared with signals collected from laboratory sources to confirm the presence of a defect. Correlation plots of two AE parameters such as amplitude and counts can be compared to determine the origin of a located source. This requires hit-based systems with feature analysis capability. Different types of analysis can be carried out, depending on the aim of the test and the type of data that have been recorded. Different AE systems are generally equipped with their own specific software, although in some cases it might be possible to extract the raw data for analysis. Analysis of collected data is primarily based on located events as previously demonstrated. If a defect is being cyclically loaded, then during each loading stage stress waves will be released. These signals (stress waves) are then detected and located causing a build-up of signals from a specific area under each successive load cycle. This results in a clustering of data giving the first indication of an active defect. The location and the accuracy with which AE sources are identified depend on the type of signals that a system can record. The arrival time of higher amplitude signals is easier to identify and thus increases the precision of the source location. The presence of existing cracks decreases the accuracy as it increases the wave attenuation due to scattering and inelastic absorption compared to initially undamaged concrete. Care should be taken that reflections from welded sections, bearings, corner sections and end sections are not recorded. Reflections can cause problems with the interpretation of waveform signals and errors in recorded AE parameters. AE is a very versatile tool but it is not recommended to be relied upon solely for structural monitoring. The use of corroborative, conventional monitoring is necessary to give a degree of confidence to the AE investigation and report. AE data should be compared with parametric data logged simultaneously by the AE system to establish meaning and significance. AN.3.6-I/5

258 APPENDIX II: CASE STUDIES Contents Case Study 1 Case Study 2 Case Study 3 Case Study 4 Case Study 5 Case Study 6 Case Study 7 Case Study 8 Case Study 9 Detection of Wire Fractures Condition Assessment of Hinge Joints 1 Condition Assessment of Hinge Joints 2 Condition Assessment of Concrete Half-Joints Relaxation Ratio Analysis of RC Beams b-value Analysis of a Reinforced Concrete Beam Non-Destructive Testing of Shear Studs Fatigue Testing of a Steel Box Girder Bridge 1 Fatigue Testing of a Steel Box Girder Bridge 2 AN.3.6-AII/1

259 Case Study 1: Detection of Wire Fractures AII.1 An acoustic monitoring system was used to monitor wire fractures on a post-tensioned concrete railway viaduct with internally grouted tendons. Description of Structure AII.2 The structure comprises six spans and carries a dual carriageway over a B road, a main line railway and part of a railway station. One of the centre spans consists of a 32m long suspended span supported on half joints at the end of two 16m long cantilevers extending from the adjacent piers. Purpose of Inspection AII.3 AII.4 A Special Inspection had indicated the presence of voids, water and chlorides in the tendon ducts. An acoustic monitoring system was installed to monitor the tendons in one of the cantilevers for wire breaks. Thirty six sensors were mounted on the structure. The probability of a tendon wire break occurring in the structure was considered to be very low so a device for creating external wire breaks was installed on the structure to check the operation of the monitoring system. The structure possessed further features that made it a good candidate for acoustic monitoring. These included: Additional structural investigations were in progress. This would be enhanced by a clear indication of the presence or absence of actively fracturing wires. The structure contained features that lent themselves to monitoring, such as difficult to inspect half joints. The structure occupies a strategic position on the network and carried a high volume of HGV traffic. Benefits of using AE monitoring AII.5 The system has been in operation since mid-1998 and has provided an excellent opportunity for confronting challenges in detecting and locating post-tensioned wire breaks in noisy environments, and establishing the protocols needed to ensure the success of long-term, continuous, unattended monitoring. The viaduct has only experienced a small number of naturally-occurring wire breaks during the eight years of monitoring, although the conditions for corrosion are present. To test the monitoring system, external wire breaks have been artificially created and detected in blind trials. References Cullington, D. W., MacNeil, D., Paulson, P. and Elliot, J (1999). Continuous acoustic monitoring of grouted post-tensioned concrete bridges. Proc. 8th Int. Conf. Structural Faults & Repair-99, Commonwealth Institute, London, July, Engineering Technics Press, ISBN AN.3.6-AII/2

260 Case Study 2: Condition Assessment of Hinge Joints 1 Introduction AII.6 AII.7 AII.8 Hinges are construction details which transmit shear from one part of the structure to another while allowing rotation to occur. They are used in the structural system to simplify design and construction, particularly where differential settlement may occur and are commonly found within bridge decks and at the tops of columns which allow for rotational movement. A typical hinge detail consists of a narrowing of the concrete section in flat slab or beam and slab decks to form a throat, through which steel reinforcement passes between the cantilever and suspended span. Hinge joints are often referred to as scissor joints as a result of the crossing steel reinforcement. There are variations to this arrangement, with differing construction sequences, reinforcement details, types of reinforcement, skews and edge details, particularly where service bays are included. A typical bridge deck hinge joint is shown in Figure AII.1. The hinge design naturally cracks the concrete at the point of rotation allowing water and chlorides to percolate down through the throat to the steel reinforcement. The only accessible parts of the joint for visual inspection are the soffit and the underside notch of the throat. Little information about the condition of the joint can be gained from this type of inspection. AE can be used to detect concrete cracking and friction of existing fracture faces during the passage of heavy vehicles which can be indicative of distress resulting from loss of rebar section and deterioration of reinforced concrete about the rotational joint. AE has been used to monitor six motorway overbridges and interchanges on the M1 and M6 in the UK during 2001 and 2002, using proprietary procedures developed from collaborative research with Cardiff University. The hinge-joints had varied maintenance histories and perceived conditions. Three hinge joints were on bridges which were part of busy interchanges, whilst the other three were local bridges between villages which carried heavy vehicles significantly less frequently. All six structures were built in the mid-1960s. Test Procedure AII.9 AII.10 Two rows of AE sensors were installed on the soffit on both sides of each joint and secured with steel brackets. A sensor installation is shown in Figure AII.2. This required an overnight closure of the carriageway and the use of a scissors lift for access. The test procedure had two parts: Firstly, monitoring was carried out under normal traffic loading over one week day between 5am and 10pm. This first test established the behaviour of the joint under normal traffic loading. Secondly, AE was recorded as a laden 32 tonne aggregate lorry crossed the joint at walking speed, the position of the vehicle also being recorded. This allowed the AE from different bridge hinge joints to be compared. The AE signals detected from both tests were attributed to friction between recently formed fracture faces and new fractures from within the hinge joint. Results AII.11 Figures AII.3 AII.6 show the AE from two interchange bridges which were of the same design, skew, age and had similar loading. Both bridges were heavily used but the first (M6 J10 interchange north bridge) was considered to be in poor condition and the second (M1 J23 interchange south bridge) had been well maintained and was considered to be in good condition. AN.3.6-AII/3

261 AII.12 AII.13 AII.14 Figures AII.3 and AII.4 show the AE recorded under normal traffic loading over just one hour of active traffic flow, with the location of emissions shown over the width of the joint (vertical axis). They show a significantly higher number and energy of AE signals from the bridge considered to be in poor condition. The AE was attributed to new concrete microfracture and existing crack face fretting. Figures AII.5 and AII.6 show the AE recorded as the 32 tonne aggregate lorry passed over the joint. Figure AII.5 shows a large amount of AE activity from the loaded carriageway (particularly under the outer wheel position) from the hinge joint deemed to be in poor condition. This information could be used to guide intrusive coring or further inspection. During normal traffic loading, it was observed that even vans passing over the hinges in this bridge caused significant emission. Figure AII.6 shows that there was very little AE from the bridge perceived to be in a good condition. This suggests that the joint did not crack significantly as the load passed over it and does not have significant recent pre-existing fractures that could fret under heavy loading. A grading system was used to compare the different hinges using the AE recorded under both the loading under normal traffic and the controlled load test. The results for the six hinges tested are shown in Figure AII.7. The M6 J10 interchange bridges were graded D and E which relate to intense and very intense emission respectively. The other bridges were significantly less active. Verification of Results AII.15 AII.16 X-ray testing at the M6 J10 interchange indicated evidence of loss of section in the reinforcement, thus, confirming the poor condition of the joint. A section of the hinge joint at the north bridge at Junction 23A on the M1 (M1 J23A) was broken out following AE testing. Visual examination of the main scissors reinforcement bar found no deterioration and no evidence of corrosion. Examination of concrete also revealed no in-service cracking. This correlated well with the insignificant A grade obtained from AE results. AN.3.6-AII/4

262 Figure AII.1: Typical Bridge Deck Hinge Joint Figure AII.2: An AE Sensor Mounted Next to a Hinge Joint AN.3.6-AII/5

263 Figure AII.3: AE Activity vs. Time Over the Width of M6 J10 Joint AN.3.6-AII/6

264 Figure AII.4: AE Activity vs. Time Over the Width of M1 J23 Joint Figure AII.5: AE Activity on South Carriageway Under Load Test, M6 J10 AN.3.6-AII/7

265 Figure AII.6: AE Activity on South Carriageway Under Load Test, M1 J23 Figure AII.7: Hinge Joint Grading Table and Results from Six Joints AN.3.6-AII/8

266 Figure AII.8: Hinge Joint Composite Rating AN.3.6-AII/9

267 Case Study 3: Condition Assessment of Hinge Joints 2 AII.17 AII.18 AII.19 A single carriageway concrete bridge spanning the M4 in South Wales was monitored for AE. The deck comprised a slab containing hinges and was suffering from spalling of the concrete on the soffit. Sensors were mounted on the soffit and held in position with aluminium clamps. AE was monitored for three days under normal traffic loading. It was found that all vehicles caused emissions. Heavy vehicles caused a larger number of emissions than cars, and heavy vehicles moving quickly across the joint caused larger energy emissions. Figure AII.9 shows the planar location results from three days of monitoring. Distinct bands of emission can be seen which coincide with the position of the internal hinge joint reinforcement. These bands occur at larger intervals than the 1ft distance shown on the construction drawings. This implies that AE is detecting sources at some bars but not every bar. It is believed that the detected activity is due to regions of concrete micro-cracking around the reinforcing steel. Figure AII.9: Location Results from three days of monitoring AN.3.6-AII/10

268 Case Study 4: Condition of Assessment of Concrete Half-Joints Introduction AII.20 NDT Techniques Half-joints are stepped joints in beams and slabs which provide for rotational and sometimes longitudinal movement at the half-depth of the section. It is difficult to keep these joints watertight and they often suffer from chloride attack, reinforcement corrosion and delamination. By their nature they are inaccessible for inspection, testing or repair, leading to difficulties in assessment of their strength in their deteriorated state and in subsequent management of their condition. AII.21 Attention, therefore, turns to NDT techniques, but radar, impact-echo and ultrasonics all run into difficulty due either to the distance to the joint at mid- depth of the member, which makes it difficult to reveal reinforcement details, or to the air-gap of the joint itself which makes it difficult to detect delamination. Radiography has proved useful for half-joints but the complexity of the reinforcement means that bar profiles which would reveal pitting corrosion can sometimes be masked by the presence of other bars. Acoustic Emission AII.22 Acoustic emission can detect concrete cracking during the passage of heavy vehicles which can be indicative of distress resulting from loss of rebar section. A trial was carried out for the Highways Agency on half-joints on Borrowbeck Viaduct on the M6 in Cumbria while it was scaffolded for repairs. See Figure AII.10. The half-joints are post-tensioned and contain horizontal and diagonal reinforcing bars. See Figure AII.11. They suffered from chloride attack and delamination. Procedure AII.23 A single AE data sensor for zonal assessment was mounted on the soffit of each of eight halfjoints whose suspected condition ranged from good through cracked to visibly corroded. The sensor chosen was capable of detecting a pencil lead break at a distance of 6m, i.e. able to detect signals of microfracture in concrete with energies of Joules 10,000,000 times smaller than post-tensioned wire fractures. The displacement of the half-joint soffit was monitored throughout the test and logged with the AE data, together with temperature. Prioritisation AII.24 The joint which was seen to be corroded emitted 67% of the cumulative energy of the eight joints, while the four known to be cracked emitted 10%, 10%, 5% and 2% of the cumulative energy. See Figure AII.12. The three thought to be in good condition emitted 3%, 2% and 0.2% of the cumulative energy. The emission from the corroded half-joint was directly related to displacement under live load. Signal analysis after the removal of extraneous noise indicated that the signals were from micro-fracture of the concrete and from secondary crack closure. Comparison of the data allowed condition ranking of the half-joints to facilitate the prioritisation of the half-joints for further AE monitoring. AN.3.6-AII/11

269 Detailed Monitoring AII.25 Eight sensors were then mounted on the corroded half-joint, three on the soffit, three on the external side of the joint and two on the vertical end face. See Figure AII.13. AE signals from the beam and the half-joint bearing were filtered out and the locations of the emissions were calculated by triangulation from the time arrival of the signals. The majority of the emissions appeared to originate from two locations, one at an area of high shear stress within the nib, and the other from a delaminated area on the side of the half- joint nib. See Figure AII.14. Most of the activity occurred at maximum live load and indicated active micro-fracture in the concrete. The source within the nib was close to a post-tensioned anchorage, and it was not therefore clear whether the micro-cracking was associated with a problem related to the anchor or to a problem associated with corrosion of the rebars. References Watson, J.R., Cole, P.T., Kennedy-Reid, I. and Halliday, J., (2002). Condition Assessment of Concrete Half Joints IABMAS (2002) 14-17th July (2002) Barcelona Figure AII.10 Borrowbeck Viaduct AN.3.6-AII/12

270 Figure AII.11 Borrowbeck Half-Joints Figure AII.12 Half-Joint Prioritisation from AE Signals AN.3.6-AII/13

271 Figure AII.13 AE Sensors on Half-Joint AN.3.6-AII/14

272 Figure AII.14 AE Signals from Half-Joint AN.3.6-AII/15

273 Case Study 5: Relaxation Ratio Analysis of RC Beams AII.26 AII.27 Acoustic emission from a number of reinforced concrete beams was recorded as they were loaded to failure. The beams were simply supported and load was applied at two points using hydraulic jacks. The load positions were varied for the different specimens according to the type of failure that was required and the loading was applied in cycles of varying steps. A complete cycle consisted of four sequential phases: loading, gauge reading at constant load, crack mapping at constant load, and unloading. The beams were monitored by an AE system using a varying number, location and type of sensors. Some of the beams were tested at the University of Edinburgh, Scotland and some at Kumamoto University, Japan (Colombo, Forde, Main & Shigeishi, 2005). The data from several beams were analysed to determine the relaxation ratio (see below). The beams chosen for the analysis covered a variety of failure modes, designs, load configurations, concrete properties and types of sensor. Relaxation Ratio Analysis and Results AII.28 AII.29 It was observed that during the early cycles (at low loads) no or very low AE activity was recorded as the load was removed from the specimens but as the load was increased the AE activity during unloading increased. In the light of these observations, a relaxation ratio was proposed to quantify and compare AE activity during the loading and unloading phases. The relaxation ratio was expressed in terms of energy and defined as: Relaxation ratio = average energy during unloading phase/average energy during loading phase where the average energy was calculated as the cumulative energy recorded for each phase divided by the number of recorded AE events. A relaxation ratio less than 1 implies that the average energy recorded during the loading cycle is higher than the average energy recorded during the corresponding unloading cycle, therefore, the loading is dominant. When the relaxation ratio is greater than 1 relaxation is dominant. AII.30 The analysis consisted of calculating the relaxation ratio for each loading cycle on each sample using data from all the active sensors. The results showed that the relaxation ratio was influenced by the concrete properties and by the rate of loading. Figure AII.15 shows the results obtained from some of the specimens. Initially, the loading phase is dominant and all the values of the relaxation ratio lie below the horizontal red line. An inversion of trend occurred when the load reached approximately 45% of the failure load. AN.3.6-AII/16

274 Figure AII.15 Relaxation Ratio Results. The red line indicates a ratio equal to one AII.31 The general effect can be explained as a dominance of the primary AE activity during the early stages of the fracture process when the cracks are forming and thus, the damage is still restricted. Conversely, once the damage has seriously progressed, the secondary AE activity due to the friction of the existing cracks starts to prevail manifesting itself as the beams are unloaded. This indicates that during the loading phase, the cracking sources prevail, whilst during the unloading phase, the friction sources are dominant. AII.32 The values of the relaxation ratio appeared to be related to the percentage of failure load reached in a specific cycle and are, therefore, related to the degree of damage of the beam. A value greater than 1 is indicative of dominance of relaxation phase and therefore, of structural damage. In the first data set of tests (shown above in Figure AII.15), the value of the relaxation ratio always became greater than 1 when approximately 45% of the ultimate bending load was reached. If the relaxation ratio was less than 1 it indicated that, in this particular instance, greater than 55% of the ultimate load carrying capacity remained. This gave rise to the possibility of using this method of analysis to predict the failure load of RC beams. However, later experiments indicated that the results were affected by the concrete strength and loading rate used during the experiments. Further calibration will be needed by the engineer. Note: These findings are part of on-going research and the use of AE for this type of application should be treated with caution. References Colombo, S., Forde, M.C., Main, I.G. & Shigeishi, M. (2005). Predicting the ultimate bending capacity of concrete beams from the relaxation ratio analysis of AE signals, Construction & Building Materials. Vol 10 Issue AN.3.6-AII/17

275 Case Study 6: b-value Analysis of Reinforced Concrete Beam AII.34 AII.35 AII.36 Acoustic emission was monitored during a load test on a 2.16m long reinforced concrete beam. The beam was of rectangular section with a width of 125mm and a depth of 270mm, and reinforced with a 16mm diameter deformed steel bar. It was simply supported, using rollers, over a span of 2m and loaded in four- point bending with the load applied in 5kN increments. The beam was monitored using a Physical Acoustic Corporation (PAC) Mistras system, in conjunction with eight PAC R6I sensors (Resonant 60kHz Integral sensor). The AE threshold was set to 35 db just above the background noise level. A b-value analysis was undertaken after the raw AE raw data had been processed. A b-value analysis uses the following modified Gutenberg and Ritcher magnitude-frequency relationship: (1) where A db is the peak-amplitude of the AE events in decibels, N is the incremental frequency which is arrived at after some preliminary tests and a and b are empirical constants. The b-value is obtained by dividing the amplitude range into a number of steps and plotting the log of the number of events within each amplitude range against frequency. The b-value is then calculated from the slope of the line through the data. AII.37 AII.38 In general terms, when distributed microcracks are developing during the early stages of damage, the b-value is high and when the macrocracks begin to localise the b-value is low. To determine the b-value for the tests described above, the events recorded during a loading cycle were divided into groups of a hundred, i.e. the first hundred events recorded were placed in the first group, the second hundred events in a second group and so on. The groups were then further sub-divided into 5 db amplitude ranges starting at 35 db and the number of events in each amplitude range (N) was determined and plotted against amplitude (A db ) and a straight line fitted using the least squares method to give the b-values. The b-values were then plotted against time for each cycle of the test and for each channel to give the change in b-value as the test progressed. AII.39 A clear pattern emerged during the early cycles when the microcracks were forming with the b- value falling as the load increased. However, once the externally visible macrocracks were formed (the beam was failing) there were fewer and more scattered AE sources and the pattern was less clear. AII.40 AII.41 The trend of the b-value for one of the channels during the second cycle of the test is shown in Figure AII.16. The load is indicated by the dotted line with a scale shown on the vertical axis on the right and the error bars show the range of b-values obtained. The localisation predicted on the basis of the minimum b- value is shown by the arrow. Notwithstanding the presence of fluctuations in the trend (due to the presence of reinforcing bars in the concrete), it is possible to see a pattern showing a decrease in the b-value at the beginning of the test (the dashed line in the chart) when the load is increasing, and then a transient during the relaxation when the load is held constant. The decrease in the b-value occurs as the beam is first loaded when cracks are being initiated. The b-value reaches a minimum when the load and the damage to the beam are at their maximum. Its value can, therefore, be used to provide a good indication of the damage that is occurring in the beam during the early stages of failure, when the beam is damaged but still structurally intact. This analysis could be used to identify internal damage within a beam, before the external cracks are visible. AN.3.6-AII/18

276 Note AII.42 The investigation showed that a b-value analysis is more useful on sound structures as it provides information from when microcracking occurs to the point where macro fractures occur by localisation (i.e.: the number of microcracks is high and they join, thereby creating a localised macrocrack). Figure AII.16 b-value Over Time for One Channel During the Second Load Cycle AN.3.6-AII/19

277 Case Study 7: Non-Destructive Testing of Shear Studs AII.43 AII.44 AII.45 AII.46 AII.47 AE monitoring trials were undertaken in the UK on steel box girder bridges in South Wales and on the M6 near Walsall. AE activity was detected in the top flanges possibly emanating from the shear studs. These results led to a two year research and laboratory testing programme at Cardiff University which confirmed that AE can be used to detect and locate fatigue damage in the shear studs. Figure AII.17 shows a laboratory specimen used to investigate the use of AE to identify fatigue cracking of shear studs in composite beams. Twelve shear studs were welded to an I-beam, three of which were notched (shown solid) prior to being encased in concrete. AE sensors were mounted to the web of the I-beam, which could be the upper flange in the case of a box-girder. The specimen was cyclically loaded in three-point bending until the composite action between concrete and the steel beam failed. Figure AII.18 shows the location results from the region surrounding the left hand stud during the early stages of the test. It is evident there is a defined peak close to the location of the shear stud and shows that AE is detecting the failure of the shear stud. Once composite action had failed, the concrete was carefully removed. A number of shear studs had become completely detached from the specimen. The results of an ultrasonic C-scan investigation on the two right hand shear studs are shown in Figure AII.19. AE detected damage in both these studs with significantly higher energy level signals detected in the stud with a 25% through crack. AE from the undamaged stud was identified as due to friction between the stud and the concrete. A preliminary field investigation based on this research programme was completed on a bridge in The investigation included the coring of a number of studs and ultrasonic testing (UT). Many of the shear stud welds cored were found to be of a poor quality with pronounced heat affected zone (HAZ) and a high degree of porosity. These flaws were located in the position where fatigue detects are commonly found and interpretation of UT results was difficult. Comparison of the physical condition of studs with initial AE results from field tests using the developed technique showed the AE procedure to be oversensitive in field investigations. From this valuable correlation experience, early test procedures in AE have been updated and further improved. Figure AII.17 Laboratory Specimen AN.3.6-AII/20

278 Figure AII.18 Failure of Shear Stud During Laboratory Investigation Figure AII.19 C-scan of Shear Studs (a) Showing Cracking 25% into the Cross-Sectional Weld Area (b) Crack Initiation of Shear stud with small Weld Flaw AN.3.6-AII/21

279 Case Study 8: Fatigue Testing of a Steel Box Girder Bridge 1 AII.48 AII.49 AII.50 Figure AII.20 shows two planar location plots recorded during global and local investigations on a bridge. The global plot shows a region of clustered activity which coincided with a sealing weld at the end of an internal diagonal bracing member. Once the region of activity had been identified, a further smaller array of sensors was added around the suspected fault. The result of this further investigation is shown in the adjacent figure and clearly shows a clustered area of activity. When comparing the two approaches it is evident the local monitoring method provided an improved resolution of location. Using this local approach it was possible to size the fatigue crack as approximately 250mm. In this instance, the crack was growing in two directions, resulting in AE from the crack tip, and the faces were also fretting (rubbing), resulting in emission along the crack length. During this investigation a surface-mounted strain gauge was used to monitor traffic loading. Figure AII.21 shows the recorded strain over a short period during the investigation. The peak strains are associated with vehicles moving across the structure. Figure AII.22 shows a history of recorded events from the cracked region for the same time period. It is evident that the strain readings show good correlation with the detected and located acoustic emission signals. References Watson, J.R., Holford, K.M., Davies, A.W., and Cole, P.T. (2000). Boxmap non-invasive detection of cracks in steel box girders. Bridge Management Four. Thomas Telford. ISBN: Figure AII.20 Example of Global and Local Location of Same Source AN.3.6-AII/22

280 Figure AII.21 History of Recorded Events for Located Source Figure AII.22 History of Recorded Strain AN.3.6-AII/23

281 Case Study 9: Fatigue Testing of a Steel Box Girder Bridge 2 AII.51 AII.52 AII.53 AII.54 Figure AII.23 presents an example of a linear location plot from a steel box girder bridge. The data was collected from a viaduct under normal traffic loading as part of a global investigation. The boxed numbers on the plot represent the position of the sensors installed. The dashed blue lines represent the position of transverse stiffeners whilst the green chain dotted line is the location of the spliced joint. It can be seen from the plot that there are peaks of located activity along the girder. If a fault is loaded, it will emit acoustic emission from any number of processes including plastic deformation at the crack tip, crack growth and fretting. As vehicles load the bridge, AE will be released from the fault region. This will build up during repetitive cycles and provide regions of located peak activity. The plot shows a number of peaks close to the transverse stiffeners and spliced joints, and these are probably regions of fatigue crack activity. The plot also shows the limitations of linear location as the vertical location of the fault cannot be determined. However, after identifying stiffeners along the structure that are generating AE, a local planar monitoring approach can be used. Figure AII.24 shows a local planar location of a transverse stiffener identified by the linear location method. The plot is a colour intensity graph of located signals; a key is shown on the right side of the plot. The green boxed numbers represent the position of the sensors. A drawing of the stiffener has been superimposed on the plot; it shows a circular cut-out out in the middle of the stiffener and triangular sections removed from the corners. It can be seen that there is region of clustered activity in the lower left region of the plot which is due to a fatigue crack growing out from the central cut-out section. This demonstrates that linear location can be applied to bridges, however, detected regions do require further local investigations using a planar location approach. References 1. Carter, D., Pullin, R. and Holford, K. (1998), 23rd European conference on Acoustic Emission Testing. Venice, Italy. Figure AII.23 Example of Linear Location from Bridge Investigation AN.3.6-AII/24

282 Figure AII.24 Example of Planar Location from Bridge Investigation Acknowledgements The Highways Agency is indebted to Pure Technologies Ltd and to Physical Acoustics Ltd for their assistance in developing the Case Studies and for the provision of illustrations. AN.3.6-AII/25

283 ENQUIRIES 7.1 All technical enquiries or comments on this document or any of the documents listed as forming part of the NRA DMRB should be sent by to addressed to the following: Head of Network Management, Engineering Standards & Research National Roads Authority St Martin s House Waterloo Road Dublin 4 Pat Maher Head of Network Management, Engineering Standards & Research AN

284 Ionad Ghnó Gheata na Páirce, Stráid Gheata na Páirce, Baile Átha Cliath 8, Éire Parkgate Business Centre, Parkgate Street, Dublin 8, Ireland (01) (01)